Is it safe to go back to work?

A blog about COVID 19 in Malaysia.   Updated 9 Aug 2020:   

It may seem academic today, but on 1 May 2020,  Malaysia had about 6,000 cases in total.  The week before there was an average of 57 daily cases.  So there was real concern whether the virus had been brought under control.  The comparison between Malaysia and the US was to give some perspective to this concern.    Although the article was first published on 1st May 2020, the analysis and conclusions still remain valid.  In fact today, the US has a higher number of infections making the comparative numbers bigger.  

Designed by Freepik

It looks as if the MCO is going to be relaxed slowly even if we still have double digits daily new cases. Is it safe to go back to work?

In the US, the message is that there must be sufficient testing and/or testing capacity before opening up the economy. With only 0.5 % of Malaysians tested so far, what confidence do we have that all is safe?

I would like to provide some data-driven rationale on why the situation in Malaysia is different from that in the USA and that we can safely proceed with opening up the economy as follows: 

1) The general public in Malaysia have less chance to come into contact with those infected as the Malaysian protocol is to hospitalize all COVID 19 cases whereas, in the US, the mild cases self-quarantine at home and only the serious cases are hospitalized. 

2) It is 105 times more likely for a person in the USA to come into contact with a mild or asymptomatic person with COVID 19 than for a person in Malaysia. 

This was estimated using a backward-looking mortality analysis which showed that the mild and/or asymptomatic cases in the USA is equivalent to 57 cases per 1,000 population whereas for Malaysia it is 0.5 cases per 1,000 population. (Refer to Note 1)

The higher deaths in the US suggests that about 6 % of the population has been infected whereas for Malaysia it is around 0.07%. 

3) Malaysians have a better chance of knowing where or who to avoid as about 3/4 of the cases in Malaysia are from identifiable groups (clusters, EMCO areas, returnees) leaving the balance as community spread. 

My view is that the critical number to observe is the one from the community spread as by definition we don't know where or who the source is. There has not been much news in the US about clusters and the news seems to suggest that the majority are from community spread.

Over the past week, we estimated that the number of cases from community spread averaged about 14 cases per day. (refer to Note 2). To put this in perspective prior to 12 Mac (which we can safely conclude that the cases were all from the 1st wave ie pre-Sri Petaling), the highest number of daily cases was 28.

4) In terms of testing, Malaysia has done better than the US. This is based on the positivity rate (defined as the number of positive cases / total number of tests) where Malaysia 3.8% is much lower c/w 17.5 % for the US. 

A high positivity rate implies that only those with a very high chance of being infected are tested. WHO standard is for < 10 % detection rate.

The figure for the US meant that the testing is not sufficient. (Refer to Note 3). Note that the number of tests per capita is not as reliable an indicator of testing sufficiency as detection rates.

Conclusion

While I think that although Malaysia is not as aggressive as the US in ramping up the number of tests, I feel much more confident looking at the above. 

While I may not have 100% accuracy in the figures since my data were from publicly available information, I think it provides an appropriate order of magnitude. 

 
With a Federal government in control of public health, Malaysia will be in a better position to have a coordinated response to any surge as compared to the US where each state is responsible for its own public health. We have EMCOed areas and I think the US will have difficulty trying this.

Note
1. The backward-looking estimates of the number of infections are based on the number of deaths divided by the mortality rates and as it takes some time from infection to death, the computed number for today actually represents the number of cases some time ago. 

In our analysis, we have assumed a mortality rate of 0.5%, and the time from infection to death is 2 weeks. 

We carried out a 6 weeks analysis based on the number of deaths in Malaysia and the USA from 12 Mac to 29 Apr to represent the number of infections from 27 Feb to 15 April (the reference date).

USA - With 58,327 deaths for the period we estimated that there are 11.665 million cases c/w the reported 609,516 on 15 April ie there are 19.14 more cases than that reported. Using this ratio and with cum 1.036 million cases on 28 April, we projected that there are actually cum 19.828 million cases. 

Assuming that all the reported cases were hospitalized, this meant that there were 19.828 – 1.036 = 18.792 million cases that were either mild or asymptomatic, equivalent to 57 cases per 1,000 population

Malaysia – With 100 deaths for the period we estimated that there are 20,000 cases c/w the reported 5,072 on 15 April ie there are 3.94 more cases than that reported. 

Using this ratio and with 5,851 cum cases on 28 April, we projected that there are actually cum 23,072 cases. Assuming that all the reported cases were hospitalized, this meant that there were 23,072 – 5,851 = 17,221 cases that were either mild or asymptomatic, equivalent to 0.5 cases per 1,000 population

This analysis is to provide a comparative order of magnitude and is likely to underestimate the true number of cases as using this methodology for Sweden, we found that 8 % of the population has been infected. 

However, the epidemiologist responsible for Sweden's COVID 19 strategy was reported to have said that about 25% of the population has already been infected. This could mean that either the mortality is lower, or the time from infection to death is longer, or that we have not sufficiently accounted for the cases for the period between the reference date and currently or all of these. 

The analysis indicates that a high proportion of those infected is asymptomatic - 75 % for Malaysia and 95 % for the US. 

This is consistent with studies of cases in Iceland and Vo where the whole population/community is tested. However, 95 % for the US looks high may suggest so doubt about the US data.

Note that a sensitivity analysis showed that if the mortality rate is increased, the number of times for a person in the US to come into contact with an infected period c/w Malaysia increases.


2. We broke down the MOH reported daily new cases into the following.

Date - April	Total	Returnees	Cluster/EMCO	Bal = community spread
	No of cases per day				Bal % of total
24	88	13	62	13	15
25	51		32	19	37
26	38		34	4	11
27	40
28	31	10	10	11	35
29	94	72	12	10	11
30	57	25	7	25	44
Average				14	26

Note that the info about the breakdown had to be deduced from the various media reports and there were days like 27 April where we could not find any info

3. As more people are tested the detection rate would come down.

     Date        US       Malaysia
                  Detection rate (%)

     5 April      19.1      7.0

    22 April    19.5       5.1

    29 April    17.5      3.8

- - - - - - - - - - - - - - - - - - - - - - - - - - -

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.