The recent announcement about the vaccine for Covid-19 led to the
following question being included in the readers' comments section of a
UK news-site report

"Someone needs to look at this 90% effective claim. Its based on
finding that 90% of those that developed Covid had been given the
placebo ... but, assuming 50% of the overall subjects were given the
placebo, then if 50% of those developing Covid had been given the
placebo the Vaccine would be 0% effective. Someone other than the PR
dept should do the maths and say what the findings actually mean."

Is there some standard way of defining the effectiveness of a medical treatment?

Note 43,500 in the total sample, 94 of whom developed Covid.

--- SoupGate-Win32 v1.05
* Origin: fsxNet Usenet Gateway (21:1/5)

On Mon, 9 Nov 2020 16:08:35 +0000 (UTC), "David Jones"
<dajhawkxx@nowherel.com> wrote:

The recent announcement about the vaccine for Covid-19 led to the
following question being included in the readers' comments section of a
UK news-site report

"Someone needs to look at this 90% effective claim. Its based on
finding that 90% of those that developed Covid had been given the
placebo ... but, assuming 50% of the overall subjects were given the
placebo, then if 50% of those developing Covid had been given the
placebo the Vaccine would be 0% effective. Someone other than the PR
dept should do the maths and say what the findings actually mean."

Is there some standard way of defining the effectiveness of a medical >treatment?

Note 43,500 in the total sample, 94 of whom developed Covid.

I assume that they were using the natural-language version of
effectiveness and failed to spell out the details. How many cases
were avoided or prevented?

Taking an "expected number of cases" as equal to Control, the
fraction of cases that were prevented comes to 8 of 9 -- 89%,
which rounds to 90%.

Is there another meaning?

--
Rich Ulrich

--- SoupGate-Win32 v1.05
* Origin: fsxNet Usenet Gateway (21:1/5)

Rich Ulrich wrote:

On Mon, 9 Nov 2020 16:08:35 +0000 (UTC), "David Jones" <dajhawkxx@nowherel.com> wrote:

The recent announcement about the vaccine for Covid-19 led to the
following question being included in the readers' comments section
of a UK news-site report

"Someone needs to look at this 90% effective claim. Its based on
finding that 90% of those that developed Covid had been given the
placebo ... but, assuming 50% of the overall subjects were given the placebo, then if 50% of those developing Covid had been given the
placebo the Vaccine would be 0% effective. Someone other than the PR
dept should do the maths and say what the findings actually mean."

Is there some standard way of defining the effectiveness of a
medical treatment?

Note 43,500 in the total sample, 94 of whom developed Covid.

I assume that they were using the natural-language version of
effectiveness and failed to spell out the details. How many cases
were avoided or prevented?

Taking an "expected number of cases" as equal to Control, the
fraction of cases that were prevented comes to 8 of 9 -- 89%,
which rounds to 90%.

Is there another meaning?

I think the problem was the apparent direct transposition of the
percentage of observed cases having the placebo into the
"effectiveness", which is more noticably wrong for a lower
effectiveness.

--- SoupGate-Win32 v1.05
* Origin: fsxNet Usenet Gateway (21:1/5)

On Mon, 9 Nov 2020 18:58:45 +0000 (UTC), "David Jones" <dajhawk18xx@@nowhere.com> wrote:

Rich Ulrich wrote:

On Mon, 9 Nov 2020 16:08:35 +0000 (UTC), "David Jones"
<dajhawkxx@nowherel.com> wrote:

The recent announcement about the vaccine for Covid-19 led to the
following question being included in the readers' comments section
of a UK news-site report

"Someone needs to look at this 90% effective claim. Its based on
finding that 90% of those that developed Covid had been given the
placebo ... but, assuming 50% of the overall subjects were given the
placebo, then if 50% of those developing Covid had been given the
placebo the Vaccine would be 0% effective. Someone other than the PR
dept should do the maths and say what the findings actually mean."

Is there some standard way of defining the effectiveness of a
medical treatment?

Note 43,500 in the total sample, 94 of whom developed Covid.

I assume that they were using the natural-language version of
effectiveness and failed to spell out the details. How many cases
were avoided or prevented?

Taking an "expected number of cases" as equal to Control, the
fraction of cases that were prevented comes to 8 of 9 -- 89%,
which rounds to 90%.

Is there another meaning?

I think the problem was the apparent direct transposition of the
percentage of observed cases having the placebo into the
"effectiveness", which is more noticably wrong for a lower
effectiveness.

Right.

That question was from readers' comments, from some news report.

I don't recall if confusion was likely in the news report that I read.
--
Rich Ulrich

--- SoupGate-Win32 v1.05
* Origin: fsxNet Usenet Gateway (21:1/5)

The method used could take account of patient-days as different people had their jabs at different times. JOHN BIBBY




On Monday, November 9, 2020 at 4:08:43 PM UTC, David Jones wrote:

The recent announcement about the vaccine for Covid-19 led to the
following question being included in the readers' comments section of a
UK news-site report

"Someone needs to look at this 90% effective claim. Its based on
finding that 90% of those that developed Covid had been given the
placebo ... but, assuming 50% of the overall subjects were given the
placebo, then if 50% of those developing Covid had been given the
placebo the Vaccine would be 0% effective. Someone other than the PR
dept should do the maths and say what the findings actually mean."

Is there some standard way of defining the effectiveness of a medical treatment?

Note 43,500 in the total sample, 94 of whom developed Covid.

--- SoupGate-Win32 v1.05
* Origin: fsxNet Usenet Gateway (21:1/5)

Rich Ulrich wrote:

On Mon, 9 Nov 2020 18:58:45 +0000 (UTC), "David Jones" <dajhawk18xx@@nowhere.com> wrote:

Rich Ulrich wrote:

On Mon, 9 Nov 2020 16:08:35 +0000 (UTC), "David Jones"
<dajhawkxx@nowherel.com> wrote:

The recent announcement about the vaccine for Covid-19 led to the
following question being included in the readers' comments

section >> > of a UK news-site report

"Someone needs to look at this 90% effective claim. Its based on
finding that 90% of those that developed Covid had been given the
placebo ... but, assuming 50% of the overall subjects were given

the >> > placebo, then if 50% of those developing Covid had been
given the >> > placebo the Vaccine would be 0% effective. Someone
other than the PR >> > dept should do the maths and say what the
findings actually mean." >> >

Is there some standard way of defining the effectiveness of a
medical treatment?

Note 43,500 in the total sample, 94 of whom developed Covid.

I assume that they were using the natural-language version of
effectiveness and failed to spell out the details. How many cases
were avoided or prevented?

Taking an "expected number of cases" as equal to Control, the
fraction of cases that were prevented comes to 8 of 9 -- 89%,
which rounds to 90%.

Is there another meaning?

I think the problem was the apparent direct transposition of the
percentage of observed cases having the placebo into the
"effectiveness", which is more noticably wrong for a lower
effectiveness.

Right.

That question was from readers' comments, from some news report.

I don't recall if confusion was likely in the news report that I read.

There is a little more information about the trial results here ... https://phastar.com/blog/250-statisticians-view-on-pfizer-covid19-vaccine-data

--- SoupGate-Win32 v1.05
* Origin: fsxNet Usenet Gateway (21:1/5)

On Mon, 9 Nov 2020 16:08:35 +0000 (UTC), "David Jones"
<dajhawkxx@nowherel.com> wrote:

The recent announcement about the vaccine for Covid-19 led to the
following question being included in the readers' comments section of a
UK news-site report

"Someone needs to look at this 90% effective claim. Its based on
finding that 90% of those that developed Covid had been given the
placebo ... but, assuming 50% of the overall subjects were given the
placebo, then if 50% of those developing Covid had been given the
placebo the Vaccine would be 0% effective. Someone other than the PR
dept should do the maths and say what the findings actually mean."

Is there some standard way of defining the effectiveness of a medical >treatment?

Note 43,500 in the total sample, 94 of whom developed Covid.

This week, there is a newer report with even better outcomes.
Only 5 of 95 cases were from the vaccine group. There were 11
"severe" cases and all were in the placebo group.

The simplest way to avoid inviting confusion might be to
report the result with a decimal point, which this later study
did. I think that the place after the decimal is pretty meaningless
on its own, but reporting it keeps the general "technical" level and
confusion level higher; and thus it discourages "thinking" and thus
avoids presumptious mistakes.


--
Rich Ulrich

--- SoupGate-Win32 v1.05
* Origin: fsxNet Usenet Gateway (21:1/5)