Like many of us, I have also been keenly following the numbers on the Covid 19 Pandemic. Since we live here, the focus is more on India and what is going to happen. I had developed a model to forecast by when we would start seeing a decline in deaths in India that I shared with family and friends.
Rather reluctantly, on their urging, I am sharing the findings on my blog. Please note that this is strictly an amateur's work. If the predictions don't pan out, well it will be no worse than some of the astrologers who are now running for cover!
Here goes:
Rather reluctantly, on their urging, I am sharing the findings on my blog. Please note that this is strictly an amateur's work. If the predictions don't pan out, well it will be no worse than some of the astrologers who are now running for cover!
Here goes:
- Here is where I stick my neck out. I have done an international per capita model of cases and deaths and used it to forecast what will happen in India
- Broadly there are 3 types of countries and progression of Covid 19. Asia including Turkey and Iran where the disease progressed more rapidly and counties like Malaysia and Thailand which have hardly been touched at all.
- Europe and North America which are broadly similar
- Brazil is still to stabilize
- I have taken the dates on which a decline started of cases and deaths for each country and worked out the per capita cases and deaths on that day
- This analysis results in the conclusion that India is likely to behave like West Asia - tropical counties of Malaysia, Indonesia and Thailand will do better than India.
- Deaths are more reliable than cases as a measure for forecasting as cases are dependent on how much testing is going on
- Based on the above analysis, my prediction is that India deaths/day will start declining around 1 July to 10 July. The number of deaths in India (based on a 7 day moving average are doubling in 12 days) will reach the projected number of 35,000 - 62,000 after which they should start declining as per the model.
- The decline will start earlier in places like Ahmedabad, Pune and Mumbai where the spread is already greater. However, as a whole the country will start seeing the decline on the dates stated above.
- This model assumes that the per capita infections and deaths when they reach a certain level in the population the infection progress recedes. It must be something similar to herd immunity but the pundits are reluctant to use that word
- The more you ease off from the lockdown, the quicker we will reach the tipping point. The doubling rate will become quicker.
Regards and stay safe.
Wow! Well researched Debu. Well argued and calculated forecast. Surely there are many variables that can change the outcome but this is so logical. Thanks.
ReplyDeleteVery good analysis!! Based on the now available, a tonne of data. And if you've got it wrong - you really don't need to join the astrologers!! This qualifies you to join the many experts who perhaps spoke too early and firmly 😊
ReplyDeleteSo does this indicate that while the 1st lockdown was justified in terms of messaging and preparedness (sop's, infra etc), after that we should have opened up and reached this tipping point earlier. Thereby lessening the economic and human issues that have resulted from the prolonged lockdown ??
Thanks! There first lockdown was very much required. Thereafter, I think we are seeing a diminishing utility. Also, as you say, the economic and human issues are becoming very severe. If this model is to be believed, what you can do is to delay the inevitable but eventually the numbers will catch up. I think we now need to live with the virus while protecting the more vulnerable segments.
DeleteLet us hope it is true Even if it peaks by Mid July it is a relief
ReplyDeleteThanks for your forecast
Ashok Sultan
Thanks! Let's hope so.
DeleteHope it declined before that
ReplyDeleteI hope so too but I doubt it based on international experience
DeleteYour projection is based on some very sound analysis and hypotheses hence comes across as logical and plausible. I only wonder if the large-scale reverse migration from large metros into villages may shift the needle somewhat... All in all, a very insightful read. Thanks for sharing!!
ReplyDeleteThanks! The migration will have no impact on the eventual outcome. The infection may spread faster, that's all. When it reaches a certain point it should be self limiting. That is the experience in other countries and there is no reason to believe that it will be very different in India.
DeleteThanks for your analysis. The predicted date also coincides with onset of Monsoon in most part of India. I wish Virus start withdrawing earlier.
ReplyDeleteSir, I have written one article on way to control virus. I feel that everyone is talking about social distancing and not considering the fact being away from other people is helpful but being away from new faces and avoiding population intermixing is necessary as well. Not everyone understands the mathematics behind limiting the power of exponents which is used by Virus to multiple. This article is about limiting those power to multiply by forming clusters. I saw your post in IITK alumni association and seeking your help in making my idea reach to larger audience.
ReplyDeletehttps://medium.com/@Anjaniksingh4/an-innovative-way-to-break-the-chain-post-lockdown-4838436c1f6
I read the article. It's very interesting. However, I think it's going to be difficult to change people's behaviour to the extent that you are suggesting. Perhaps we need to recognize that the virus has to run its course. In all counties the cases and deaths have started declining after a point. There is no reason to believe that it will not happen in India too.
ReplyDeleteThanks a lot for reading and providing inputs. I too agree that complete implementation is not possible but is it possible to create awareness to avoid intermixing with new faces, avoid exploring out new public resources whenever possible.
DeleteWhat reason you think behind declining active cases in most of European countries. Is it Herd Immunity?
I think that some behavioural changes like social distancing, WFH, hand washing etc will become part of life. As for the rest if there is a greater danger of infection, people will bring about the changes automatically. There is enough awareness now.
DeleteI am not an epidemiologist so I can't really say. It sounds like herd immunity but people are reluctant to use the phrase.
Thanks for your views
DeleteDear Debu: I read your blog with great interest because I belong to the group which is highest at risk: Over 70, with 3 co-morbidities. As you must know, I was in Bombay till April 9, in isolation and avoiding as much as possible contact with the invisible virus. Now I am back in my other home in Switzerland and I am seeing here how the matter is being addressed.
ReplyDeleteThe basics are the same all over the world. There are no black and white regarding this virus. Since we do not know what it does, we must rely on data as it emerges with time. Being somebody who has spent his whole career in biological/medical research, it is necessary to talk in terms of chances and risks. There are however some basics:
1. The virus will infect people till there is a vaccine or a therapeutic.
2. The chances that the virus will die on its own are very small
3. The SARS-CoV2 virus is as infectious as influenza but is at least 5-10 times as lethal.
4. People who have morbidities associated with the pulmonary and cardiovascular systems and anything which affects these systems are more at risk at any age. People older than say 65 or at the highest risk.
5. So there are 2 ways to combat this virus, a) by reducing as much as possible its transmission by lockdowns, or b) by allowing everybody except those most at risk, to get infected, recover and thus have what is well known as herd immunity.
6. Most countries have chosen the former but Sweden has chosen the later. Britain started with intention of inducing herd immunity but soon learnt that the price to pay in terms of the elderly dying was too high. Sweden is now paying that price.
7. However, only a very few countries (S. Korea, Taiwan) have analyzed the options of what one must do, completely!
8. One cannot use lockdown indefinitely, it is a first step which should prepare for the next which is identifying the infected and isolating them. To do this one must do blanket testing which and taking special precautions for the elderly which S. Korea did. They had the lowest number of deaths over time compared to any other country. The vast majority of countries have not carried out blanket testing with the consequences we see.
7. With regard to India, the start was excellent but there was no planning for the next steps. We do not have blanket testing even in the so-called "hotspots" like Bombay. Although we were able to keep the numbers very low at the beginning (just like S. Korea) we now come to a point where even with the lockdown, the areas of very high population density (like the Dharavi slums in Bombay) are going to "explode" with cases. Just as soon as the lockdown is loosened, it will be the elderly who will start to get infected (part-time household help returning)and so not only will the numbers of deaths but also the infection rate will go up., unless we tested much much more and then isolate the infected.
So I am not sure whether the number of deaths will go down significantly in the foreseeable future unless we can compensate lockdown loosening with much more testing. And from what I can read from the Indian press (I read the e-papers everyday even here in Switzerland) there seems to be little chance of testing increasing, and by testing I mean RT-PCR and not antibody testing.
Please also take my comments as those from someone who is a scientist, but certainly not with expertise in infectious diseases, virology or public health, but someone who does know how to analyze data.
Finally, I very much think that we need to hear from people like you who are not only intelligent but also have social values related to doing the best we can for our fellow citizens. So once again many thanks for your comments.
Ajay Bhatnagar in Bombay and Grindelwald
Thank you for your well considered and comprehensive response. If you analyse the data from all countries, the infections and deaths start declining at some point in time. As I mentioned in my blog, there seem to be clusters of countries that behave in a broadly similar fashion. Why does this decline happen? I would be very grateful for your views.
DeleteIf you take S. Korea as an example, there the curve flattened and declined very early. In my way of thinking what happened here was that as those infected were identified by testing and isolated, the pool of people who would be susceptible to infection becomes smaller with time. As this pool gets smaller, the number of infections decreases and the R-zero value also decreases to under 1.0. Also from their data, those who were infected and were released back into the pool did not seem to get re-infected. The curve then flattens and starts to decline as the pool of "infectables" gets smaller and smaller. Again in my opinion, the number of people infected and recovered reaches the level of 75% of the population simulating what is classically known as "herd immunity".
DeleteBecause of the huge heterogeneity in India, regions need to addressed separately. Therefore it is my belief that metro Bombay is a separate "country" as compared with say the whole state of Maharashtra, which is why I have limited my comments to the Bombay metro region.
Very interesting. I was taking the example of USA where New York was the most badly affected early on. Later the virus spread to other places but eventually cases and deaths have started to decline. Do you think India may behave similarly? In the sense, Bombay which is the most affected as of now will eventually start to decline while cases in other parts of the country keep rising. Eventually, they may also reach the inflection point and we start seeing an overall decline. My feeling is that India just does not have the resources to adopt universal testing.
DeleteBombay is trying to do their best to limit transmission but without blanket testing they will not know whom to isolate and who not. Also since testing is limited, there is no good estimate of how many cases there really are! What we are seeing today is the number confirmed through limited testing. The reality will be emergence of symptoms which then will have to be used as the discriminator for isolation. The newspapers are reporting that the BMC is expanding the availability of isolation beds at an enormous rate! I just received a notice from Willingdon Club that a 250-bed ICU unit will be deployed on our grounds. My feeling is that the number of cases will double with a doubling time of near 10-15 days. What I am also hearing is that the bastis are slowly getting empty as people are going back to their villages. What effect this will have is not known but hopefully will contribute to some reduction in cases.
DeleteI listen to Gov. Cuomo's daily press conferences and in NY city they have been able to ramp up the tests they do by quite a bit. This has resulted not only in a flattening of the curve but also on a down swing. It has meant that he has had to keep isolating and thus reducing transmission. But he also knows that he has to open up sometime sooner than later. I really hope that they are now low enough on the curve that there is no second wave as that would tax very heavily on their now very low and diminishing resources.
My personal opinion on testing (and I'm no expert) is that there's no point in just testing and testing if you can't take follow up actions (like isolation) due to lack of capability and / or resource. Dharavi has a 7L population. If through some gargantuan effort all these get tested and say 1.5L turn up covid positive; what's the govt going to do? It's a tough one, but as you remark Debu, India does not have the resources to adopt universal testing - either to do the testing or follow up meaningfully on the results of the testing. Best to focus on broad containment measures, improve health infra to deal with serious cases and of course get the economy going in a calibrated fashion.
DeleteYou are absolutely correct! However, in Bombay they are feverishly expanding their availability of isolation beds. For example the so-called South Bombay hub will be 3 sites, one at the Dome of the NSCI, the other in the parking lot of the Racecourse and the third announced as recently as yesterday will be at the Willingdon Club. Although I am not quite sure of the number, there will more than a thousand beds available.
DeleteYou are absolutely right that testing without isolation facilities is not any use. Bombay I feel is going about it the right way.
Looks good. Well your assumptions are in place but it seems to have missed out an important factor I. e from June first week there will be monsoon in major part of India and my assumption is that the number of cases will spike and only by September we can see the decline in the no of cases.
ReplyDeleteThanks! I really can't say what the onset of the monsoon will bring about. After the onset of the disease this will be the first monsoon that we will be facing. Fingers crossed!
DeleteWhile it is true that the total cases are dependent on the testing rate and hence can not be used to precisely project future cases, same thing can be said for the deaths as well.
ReplyDeleteThe death data can be reported wrongly - like in case of West Bengal, deaths with comorbidity and covid19 were classified as comorbidity deaths. Further recent change of guidlines by ICMR has mentioned that a person who is brought dead to a hospital will not be checked for Covid19 - so even if someone dies of Covid19 that number won't reflect in the data.
There has been much discussion on need for data sanctity in India - even before Covid19 - but as the things stand, we have 100s of years to go before we take data seriously and hence start reporting it accurately. Until then, everything is just a guesstimate.
You are absolutely right. I picked death rate as a little better than infections as we are just not testing enough for it to be a reliable indicator.
DeleteDivyaroop Bhatnagar is inviting you to a scheduled Zoom meeting.
ReplyDeleteTopic: Coronovirus Stats in India
Time: May 24, 2020 07:30 PM India
Join Zoom Meeting
https://us04web.zoom.us/j/79450777958?pwd=akZVT2crYXYwSkRHd0tldGFhWlVKUT09
Meeting ID: 794 5077 7958
Password: 6tr7xb
Folks, if you are interested in stats for India Coronavirus, please join the Zoom meeting tomorrow. It will be mostly Maths so if that does not appeal to you it's OK
ReplyDeleteI have updated the model based on current data. Please join the Zoom Call if interested
ReplyDeleteDivyaroop Bhatnagar is inviting you to a scheduled Zoom meeting.
Topic: Coronovirus Stats in India Update 1
Time: Jun 8, 2020 07:30 PM India
Join Zoom Meeting
https://us02web.zoom.us/j/84294511071?pwd=bktQWGo5RFNYSk9uemFqVDVidE9PZz09
Meeting ID: 842 9451 1071
Password: 070254
One tap mobile
+13462487799,,84294511071#,,1#,070254# US (Houston)
+16465588656,,84294511071#,,1#,070254# US (New York)
Dial by your location
+1 346 248 7799 US (Houston)
+1 646 558 8656 US (New York)
+1 669 900 9128 US (San Jose)
+1 253 215 8782 US (Tacoma)
+1 301 715 8592 US (Germantown)
+1 312 626 6799 US (Chicago)
Meeting ID: 842 9451 1071
Password: 070254
Find your local number: https://us02web.zoom.us/u/kfJEoZPb0