Re: What is the difference between the rate of new infections?... and incidence?...

Richard Ulrich <Rich.Ulrich@xxxxxxxxxxx> wrote in
news:707sj3530p9vpjlne1s4asivoe5892ng2l@xxxxxxx:

On Thu, 15 Nov 2007 23:58:33 -0600, David Winsemius
<doe_snot@xxxxxxxxxxx> wrote:

Richard Ulrich <Rich.Ulrich@xxxxxxxxxxx> wrote in
news:om8pj35bfsd0r0q6h1q3alia6r1tt89bep@xxxxxxx:

On Thu, 15 Nov 2007 05:21:12 -0800 (PST), Gary
<LanceGary@xxxxxxxxx> wrote:

On Nov 15, 12:57 am, the zak <don.sak...@xxxxxxxxx> wrote:
What is the difference between the rate of new infections?...
and incidence?...
snip
DW>
incidence rate = events/population at risk/time ...typically
expressed as events/person-year

Incidence can also be expressed as cumulative incidence which is
just a proportion, i.e. a percentage or fraction.

zak >
For example, in regard to human immunodeficiency virus
acquired immune deficiency syndrome.

RU> >
That's a poor "for example" because HIV is not at all typical
of other diseases. Attack rate for HIV (from infection to
disease) is approximately 100%.

DW>
Except that is not what an attack rate is. Attack rate is the
probability of infection following exposure, and for sexual
contacts, the attack rate for HIV is nowhere near 100%.

RU> I agree that I could use "attack rate" in more than one way.

I disagree. Your use is simply wrong.

RU> Above, I was specific enough to be fairly unambiguous; although
perhaps I should have said "manifest disease". Someone else
might want to say that a subject "has a disease" at a low level
when the antibodies are successfully destroying it. If there is
a preferred term for what I was describing, what is it?

Perhaps "time to progression" or "mean time to progression". The
point remains that you were doing violence to the definition of
"attack rate", a term with a specific use and meaning in epidemiology.
If the OP were asking about incidence, then bringing attack rate into
the discussion might have relevance, but not by way of discussing the
rate at which HIV infection progressed to an AIDS-defining criterion.
Doing so just confused the discussion.

You were also misleading the OP by conflating a cumulative number with
a rate. A rate, when used correctly, will have a time component in the
denominator. The fact that "attack rate" violates that rule is a
historical and regretable fact. Current coinage of terms in
epidemiology is more disciplined.

--
David Winsemius
.

Relevant Pages