Queries and doubts about new Covid-19 deaths including increasing BID cases which await clarification from Khairy Jamaluddin
I have this morning received an email from a scholar tracking Covid-19 data, reinforcing doubts and queries about the recent data about Covid-19 deaths which must be clarified by the new Health Minister, Khairy Jamaluddin.
In fact, I would suggest that Khairy should make a Ministerial statement on the issue in Parliament tomorrow.
This is the email from the scholar tracking Covid-19 data in Malaysia:
“I have a serious concern about the new method of reporting Covid-19 deaths adopted by MOH. It is potentially a highly misleading way of presenting the data.
“The graph on deaths on this website https://covidnow.moh.gov.my/ purportedly shows the number of ‘actual deaths (7d average)’. This term is patently wrong. The number associated with that term is not the 7-day average of actual deaths. Rather, the number seems to be the 7-day average of hitherto confirmed actual deaths.
“Depending on how fast MOH confirms reported deaths, the latter number can be a very substantial underestimate of the former number.
“Below is a hypothetical example, assuming that MOH can clear almost all the tasks of confirmation of reported deaths 7 days ago, and the number of completed tasks decreases as the date approaches the current date.
date |
actual number of deaths |
actual number of deaths confirmed on 19/09 |
13/09/2021 |
205 |
200 |
14/09/2021 |
212 |
190 |
15/09/2021 |
198 |
150 |
16/09/2021 |
221 |
101 |
17/09/2021 |
209 |
52 |
18/09/2021 |
207 |
21 |
19/09/2021 |
201 |
3 |
7-day average of actual deaths |
7-day average of hitherto confirmed actual deaths |
|
207.5714286 |
102.4285714 |
“As you can see, the 7-day average of hitherto confirmed actual deaths is only around 50% of the 7-day average of actual deaths.
“Thus, the 7-day average of hitherto confirmed actual deaths is not anywhere near an accurate estimate of the 7-day average of actual deaths.
“When we can’t measure the exact number, an accurate estimate is often useful. However, since the estimate here is wildly inaccurate, I don’t see any value in including that information in the graph.
“There is a claim by some MOH official that this new reporting method gives a more accurate reflection of the current trend. But this is false. I’ve seen a few occasions where the number of actual deaths on the day before is a single digit which has then grown to a double-digit figure after a few days later.
“In the hypothetical example above, if the previous week 7-day average is 200, then there is no reduction in 7-day average of deaths at all. The 7-day average of hitherto confirmed actual deaths gives a misleading picture of reduction by half. The reduction is only an indicator of MOH’s inability to complete their tasks in a timely manner.
“If some cases take weeks or even months to confirm, then the estimate is even more skewed than what is depicted in the hypothetical example above.
One might argue that despite the short-term inaccuracy of this estimate, in the long-run it provides a better picture of the actual situation. But that is not true. So long as the rate of completing the task of confirming reported deaths is constant, the graph is always inaccurate at the final segment, and that segment can include the past week or the past few weeks or even the past few months, depending on how fast MOH clears the backlog.
“When the final segment is very different from the real situation, how can one make reliable extrapolation into the near future?
“MOH should give a strict deadline for staff to clear the backlog. If the deadline is 7 days, then MOH should only report the number of deaths up to one week from now. There is no point presenting grossly inaccurate estimates of data.
“Could it be that MOH realises that it has difficulty completing the tasks of confirmation within 7 days?”
Khairy should not only clarify this scholar’s queries, he should also explain why he failed in one of his priorities to reduce the Brought-in-dead (BID) Covid-19 death cases, as he had said that 86% of the 2,417 BID cases in August were not under any healthcare provider’s care.
Why has his strategy to reduce the number of BID cases by ensuring that the public get tested once they develop Covid-19 symptoms and get medical attention has failed, as the number of BID death cases in his three weeks as Health Minister is very much more than the last three weeks of the previous Minister of Health.