A blog about COVID 19 in Malaysia. Updated 9 Aug 2020:
Although the article was first published on 2nd May 2020, the analysis and conclusions on why there will not be another wave or why we don't need another lockdown still remain valid. We still do not see data-driven comments by the various experts. With hindsight, the fact that there was no 3rd wave shows the importance of data-based opinions. Furthermore, the points made in the Letter are equally valid today.
 |
| Designed by Freepik |
The problem with all the health experts’ comments so far is that I have not seen anyone offering in-depth data-driven analysis to support their positions.
For example, although half of the past 7 days cases were local transmissions, more than 3/4 of these were from EMCO areas and/or clusters with very good chances of breaking any transmissions.
In order to rectify this, I had earlier in the week summarized all my analyses and sent it to both Malaysiakini and Star with the hope that they either publish it as a letter or start a more detailed investigative piece.
In order to rectify this, I had earlier in the week summarized all my analyses and sent it to both Malaysiakini and Star with the hope that they either publish it as a letter or start a more detailed investigative piece.
Unfortunately, none of the media picked it up so I figured that either my rationale is flawed or I am not credible enough.
The letter was to provide a data-driven rationale that irrespective of the number of daily cases, we should not extend the MCO come 12 May implying that we can relax the MCO immediately.
I attach it for you all to judge.
To come back to the concern about the conditional MCO, I guess the govt position that it will not lead to a 3rd wave is based on two key premises:
- We can still minimize the contact between those currently infected and/or PUI from the general public after 4 May
- There are no significant asymptomatic cases that are free to walk about.
Maintaining contacts
With regards to minimizing contacts, we should look at this from the 3 sources of infections.1) Those under mandatory quarantine i.e. returnees and EMCO areas – there should not be much concern about them infecting the public.
2) Clusters that are not under EMCO as the people are not centrally located e.g. Sri Petaling, Kuching medical. There are still daily new cases from this group e.g. over the past 5 days we still have 60 new cases altogether from the Sri Petaling and Church clusters.
I have the least info for this group and I hope there is some voluntary stay at home and MOH is still doing contact tracing.
3) Community and/or sporadic spread i.e. don't know where or who is the source. The way to control this is testing, contact tracing and isolation.
We did this in the first wave and I think that as long as the daily number of cases is below 30, we should be OK. The weakest link in this containment approach is contact tracing. So it was good to require restaurants to keep track of their customers.
Also downloading the MyTrace app should help. I tried downloading MyTrace to see how it works but I have not been able to key in the verification code.
There are still bugs. I also noticed that there were only 500 odd downloads. I think the govt should get serious about getting more people to download this app.
Asymptomatic cases
With regards to the asymptomatic cases, I would like to bring you back to one of the 3 mysteries I posted a few days ago i.e. mystery No 2 which asked where all the large numbers of cases as computed by a backward-looking mortality analysis went to c/w with the reported number of cases?
My hypothesis on the “missing cases” is that because our contact tracing is so slow, by the time MOH identified the cluster and tested all those related to the cluster, these mild and asymptomatic cases had already recovered and no amount of PCR test was going to detect them.
The way to check this reason is to conduct the antibody test on these clusters but until then I cannot think of another reason.
If you accept this "missing cases" hypothesis, then there is no need to worry about the asymptomatic cases.
But if the hypothesis is wrong, then expect a spike in 2 to 3 weeks’ time. But it is not all that bad as the ban on public gathering and wearing face masks (not mandatory?) are targeted at controlling the spread from asymptomatic cases.
Finally, if you agree with the above rationale that there will be minimal contact between the infected and the general public, why are the barbers and hair salons still closed?
- - - - - - - - - - - - - - - - - - - - - - - -
Letter
As the MCO comes up to 6th week, I am certain many people are wondering if there could be another potential extension come mid-May especially because double-digit daily new cases (DDDNC) are still being reported rather than the aspired single digits or even complete elimination of the virus.This rationale is built around four key points
- The DDDNC will not overwhelm our medical capacity
- The majority of cases are from clusters and/or EMCO areas
- The risk of a virus resurgence Post MCO can be managed if the correct controls are put in place and maintained
- We have met the threshold of the USA and European Union guidelines for ending the lockdown
1) Medical capacity
The original goal of the MCO was to ensure that the number of infections does not overwhelm our medical capacity.Today we have about 6,900 beds and about 1,000 ventilators set aside for COVID 19 treatment.
During the MCO period, the peak bed utilization was less than 40%. Correspondingly, the maximum number of COVID 19 patients on ventilators was 66. At the same time, we do not seem to have a critical shortage of PPE or even healthcare staff.
A back of the envelope calculations (Notes 1, 2) also shows that our medical facilities (beds and ventilators) can easily support DDDNC.
Unless we are having problems with PPE and/or healthcare staffing, the analysis suggests that insufficient medical capacity cannot be a reason to extend the MCO.
One of the reasons for the MCO was to “flatten the curve” so as not to overwhelm our medical system. Not only have we flattened the curve, but the number of cases is significantly below our medical capacity even at peak infection.
2) Clusters and EMCO areas
About 2/3 of the total cases in the country are accounted for by clusters, sub-clusters, and those in the EMCO areas.
This profile meant that a larger part of the DDDNC is due to location-based or event-based transmission making it easier for detection and control. For example, on 26 April, almost all of the 38 new cases were reported to be from clusters and EMCO areas.
The trade-off is thus not between national public health and economic impact, but rather between controlling the virus within a specified area and/or community vs the national economy. MOH has already demonstrated its ability to trace such clusters and/or to execute the EMCO to keep the virus in control.
3) Resurgence post-MCO Is manageable
A common reason why we should not lift the MCO too early is the concern about a virus resurgence. There are two counter-arguments to this reason: -a) It is likely that we will have a virus surge even if we have zero reported daily cases before lifting the MCO
b) Even if we have DDDNC by the time the MCO is lifted, the virus surge can still be controlled.
a) Likelihood of a virus surge.
This argument follows this logic - a virus surge is still likely post-MCO because there will continue to be sources of infections even excluding those currently under quarantine or under EMCO:- Unless we quarantine all contacts related to new cases at the time when the MCO is lifted, there will be some infected in the incubation stage walking freely. To identify all contacts quickly may be beyond our current contact tracing capability.
- At the MCO opening date, there will still be cases under treatment. Some of our healthcare staff have been infected through work so we can expect this to be a source of new infections.
- There will still be returnees post MCO. Although under quarantine, we already have a history of such groups contributing to the daily new case numbers.
- Since we have not carried out mass testing of the general population, there may be asymptomatic cases walking around when MCO is lifted. Based on mortality rates, the actual number of cases is estimated to be 40% higher than that reported implying that we have asymptomatic cases. (Note 3)
Given these, there is little basis for any plan to achieve zero cases before lifting the MCO.
b) Capacity to handle a new surge of Post MCO infections.
In this context, there are 2 critical questions- Can we avoid a spike?
- What is the threshold for another national MCO if the virus goes out of control so that the public and business community can see what is coming and take appropriate action?
We have been able to manage to keep the virus under control during the 1st wave. It was the second wave that was a cause of concern that lead to the MCO.
- We have increased our testing capacity
- Although MOH has yet to announce them, it is very likely that we will have many social distancing and public hygiene measures post-MCO that were not there previously
- We are likely to continue with the 14 days quarantine for all traveling into the country.
It is worth revisiting the rationale for why the MCO was first announced on 16th March 2020, when there were 125 new COVID 19 cases reported. Was there a data-driven basis to establish this trigger?
IHME, the University of Washington health research centre whose mortality projections have often been quoted by the Whitehouse, had a rule of thumb (based on the capacity of the US medical system) for lifting the lockdown based on the number of daily new cases falling below 1 case per million population. Using this, our threshold would be 32 new daily cases.
Another approach is to have the number of daily cases such that R naught (R0) = 1. MOH had around 11 April stated that RO had come down to 1 and accordingly the threshold can be 150 cases (taking the average cases for 10 – 12 April). Of course, an epidemiologist would be able to provide better estimates.
An alternative is to look at the number of daily cases that our medical system can support and factor in a margin of safety to cover exponential growth.
The analysis in Note 1 shows that based on beds as the bottleneck, we can comfortably support 490 new cases per day. Our history is that during the exponential growth stage we went from 35 daily cases to 217 daily cases within 3 weeks. Using this factor, the trigger for another MCO is 80 new cases (Note 4).
The point is that ignoring IHME, the 38 cases for 26 April and any future DDDNC should be seen in the context of the above 80 – 150 cases trigger.
The point is that ignoring IHME, the 38 cases for 26 April and any future DDDNC should be seen in the context of the above 80 – 150 cases trigger.
4) Malaysia has met the current US and European Union Guidelines for opening
While MOH has yet to unveil its guidelines for lifting the MCO, many other countries who had implemented their lockdowns earlier have already issued guidelines for ending their lockdowns. It seems unlikely that the MOH guidelines or threshold would be significantly different.In any event, Malaysia has already met both the US and European Union guidelines (Note 5, 6).
Conclusion…
When the stakes are this high, all of us need to keep ourselves better informed and ensure that when major policy decisions are taken, they are grounded on facts & reliable data.A target for the complete elimination of the virus (i.e zero cases) may be impractical given the potential cost to the economy. As this potential threat seems to be here to stay, a more pragmatic approach is to learn how to ‘live with’ and manage this risk.
We all have a responsibility to adhere to the post MCO control measures (e.g. social distancing and public hygiene measures). At the same time, the authorities need to maintain vigilant monitoring of trends (number of cases) to provide enough lead time so that the public and business community will not be surprised.
While the above provides some data-driven reasons for not extending the MCO irrespective of the number of daily cases come 12 May, more importantly, they provide the rationale for relaxing the MCO even at this stage.
We all have a responsibility to adhere to the post MCO control measures (e.g. social distancing and public hygiene measures). At the same time, the authorities need to maintain vigilant monitoring of trends (number of cases) to provide enough lead time so that the public and business community will not be surprised.
While the above provides some data-driven reasons for not extending the MCO irrespective of the number of daily cases come 12 May, more importantly, they provide the rationale for relaxing the MCO even at this stage.
Note
1.
With
6,900 beds and assuming 14 days of treatment, there is the capacity to handle 490
new cases daily, assuming all who are infected are hospitalized. Furthermore, the capacity will be much higher
if we follow that practice of many countries where the mild cases are
quarantined at home. Given that about 80% of cases are mild, the beds’ capacity
could increase fivefold.
2.
The
peak ventilator requirement was 3 % of those under treatment. Based on 1,000 ventilators, and assuming that
those on ventilators require 21 days of treatment there is the capacity to handle
about 1580 new cases daily
3.
Assuming
1.6 % mortality rate and current mortality reflects that 2 weeks ago, we
projected 5,713 cases c/w actual 4,119 ie 39 % higher for week 5 of MCO
4.
With 79 cases at the start of exponential
growth, we expect 490 cases 3 weeks later. Even if MCO is implemented
when we hit 490 cases, there would be many contacts at the incubation stage. By
week 3, there would be about 4,800 infected persons before the impact of any MCO
begins to bring the spread down. The 4,800 cases are well below our 6,900
beds’ capacity.
5.
The “opening up America again” guidelines had a “gating criteria”
covering symptoms, cases, and hospitals. Since there is no public data about how
Malaysia is tracking symptoms, we looked at performance under the other 2
criteria.
a)
Under cases: Downward trajectory of documented cases within a 14-day
period OR Downward trajectory of positive tests as a % of total tests within a
14-day period - We have met both criteria
b)
Under hospitals: Treat all patients without crisis care AND Robust
testing programme in place for at-risk healthcare workers, including emerging
antibody - Note
2 shows that we have sufficient medical capacity and MOH guidelines for healthcare staff to meet this.
6.
We
have also met the European Union roadmap for exiting lockdown based on 3
criteria
a)
Epidemiological,
the spread should be reduced for some time
b)
Sufficient
health system capacity
c)
Monitoring
capacity: testing, contact tracing and isolation
- - - - - - - - - - - - - - - - - - - - - - - - - - -
Be fearless to get more to Stay Safe, Share this Blog
PS: This blog is for me to better understand COVID 19 as this will impact my investments. If you are also into equities, follow me at i4value.asia
Disclaimer: I am not an epidemiologist, healthcare worker, pharmacist, or staff in the Ministry of Health, but rather is someone with a strong interest in numerical analysis. The content is an attempt to understand what is happening in the battle against COVID 19 from a data-based perspective. The opinions expressed here are based on information extracted from readily available public sources but I do not warrant its completeness or accuracy and should not be relied on as such.
