Covid-19 Swabs Test Analysis :Data Error?

Is our infection rate going up?

One of the ways to analyse is we are having a lower infection rate is to look at the % of confirmed cases vs the Total number of People Tested.

We did an analyse a few weeks back and deduce that the number of infection had reduced from 28.5% from week of Apr 14 to 21.8% from week of Apr 27. Today, we will look at the number from May 4 to May 18

Anaylsis

 Source: MOH

May 4
Total Swabs Test: 175603
Total Unique Test: 123525


May 10 
Total Swabs Test: 246254
Total Unique Test: 186183

Between May 4 and May 10, the number of Swabs Test was at 70 651. This is an average of 10093 tests per day. For a unique Person swabs test, it will be 62658. This averages 8951 per day. Compared to the last analysis where average was at 2900 per day, it is a significant increase.

Let's look at the number of confirmed cases during this period

May 4: 573
May 5: 632
May 6: 788
May 7: 741
May 8: 768
May 9: 753
May 10: 876

Total: 5131
Total Unique Tests: 62658
% of infection:  8.19%

Compared to the last analysis of 21.8%, this is a vast improvement in the infection rates. This gives the impression that the situation is much better than compared to the last month.

Source: MOH

Source: MOH

May 10
Total Swabs Test: 246254
Total Unique Test: 186183

May 18
Total Swabs Test: 294414
Total Unique Test: 191260

Between May 10 to May 19, the number of Swabs Test was at 48160. This is an average of 5351 tests per day. The number of unique tests was at 5077. This averages 565 tests per day. The number of unique tests is surprisingly low and might not be right.

Let's look at the number of confirmed cases during this period


May 10: 452
May 11: 486
May 12: 884
May 13: 675
May 14: 752
May 15: 793
May16: 465
May 17: 682
May 18: 305

Total: 5494
Total unique tests: 5077
% infected: 108%??

Data Error?
The numbers do not seem right. With 5077 unique tests, there are a total of 5494 confirmed cases from the same period.

Here are some possibilities

• The previous unique test number (May 10) was incorrect.

• The current unique test number (May 18) is wrong. 

• There is the possibility of Swabs test on existing non-unique cases tested positive.

• The numbers are correct and the infection rate has shot up from a low of 8.19% from the week of May 4 to over 100% from the week of May 10! (Highly unlikely)

Given the discrepancies, MOH should clarify on the numbers. It could be just an error that needs to rectify or a potentially worse scenario than what was presented. With the circuit breaker to end in less than 2 weeks, the public would need all the reassurance they can get on this.


Revise Analysis from May 4 to May 18
Given the discrepancies in the numbers, we decide to do an analysis base on May 4 and May 18 numbers instead.

May 4
Total Swabs Test: 175603
Total Unique Test: 123525

May 18
Total Swabs Test: 294414
Total Unique Test: 191260

Base on the numbers, 118811 additional swabs tests, averaging 7920 tests. Unique tests are at 67735, averaging 4515 per day.


May 4: 573
May 5: 632
May 6: 788
May 7: 741
May 8: 768
May 9: 753
May 10: 876
May 11: 486
May 12: 884
May 13: 675
May 14: 752
May 15: 793
May16: 465
May 17: 682
May 18: 305

Total: 10173

Total unique tests: 67735

Total infected: 15%

Base on this revised analysis, there are a few things we noted

1) Swabs Test and Unique Test have both increases in average

2) The infection rate is coming down from mid 20 to 15%. 

As mention in our previous post, a more accurate way to interpret infection rate would be base on the number of unique tests administer daily vs confirmed cases. It will be more representative of the trend of infection. Looking at the numbers on its own will not yield any conclusive analysis of the situation and rather pointless for that mater.