Date: Tue, 30 Apr 2002 09:53:49 -0700
Reply-To: Dale McLerran <stringplayer_2@YAHOO.COM>
Sender: "SAS(r) Discussion" <SAS-L@LISTSERV.UGA.EDU>
From: Dale McLerran <stringplayer_2@YAHOO.COM>
Subject: Re: Age calculation in epidemiology
In-Reply-To: <OFF5D233F0.DFFC0529-ON85256BAB.004811A0@kendle.com>
Content-Type: text/plain; charset=us-ascii
I basically agree with Dennis. However, if you are rounding age to
the nearest year vs counting completed anniversaries of a persons
birthdate and using that variable as a continuous predictor variable,
about the only statistic which would change would be the intercept.
The slope associated with age should be essentially the same. Thus,
I see no advantage to rounding age to the nearest year over the
traditional definition of age.
I believe that it might have been Jay Weedon who suggested that you
are not likely to get the detail about how age is computed from
journal articles. My belief is that this detail would likely be
provided if the authors employed the rounding approach. However,
if age conforms to standard definitions, then it is not necessary to
define how age is computed. The silence in the literature regarding
algorithm for computing age would be an argument that age is obtained
employing completed anniversaries.
In summary, it is my belief that the epi literature employs traditional
age definitions, and there is little reason to deviate from the
standard definition.
Dale
--- diskin.dennis@KENDLE.COM wrote:
> Dietrich,
>
> I've heard arguments for both approaches.
>
> Option 1 makes the most sense if you are going to use age as a
> continuous
> variable, as in a regression but, then why round it ?
>
> If age is going to be used as a categorical variable for grouping as
> you
> imply, then option 2 will give you results that can more easily
> compared to
> population statistics etc. Also, if you are grouping anyway, then the
> cutoffs are arbitrary so the traditional approach makes as much sense
> as
> the other.
>
> FWIW,
> Dennis Diskin
>
>
>
>
> From: Dietrich Alte <alte@UNI-GREIFSWALD.DE>@LISTSERV.UGA.EDU> on
> 04/30/2002 05:47 AM
>
> Please respond to alte@uni-greifswald.de
>
> Sent by: "SAS(r) Discussion" <SAS-L@LISTSERV.UGA.EDU>
>
>
> To: SAS-L@LISTSERV.UGA.EDU
> cc:
>
> Subject: Age calculation in epidemiology
>
>
> Dear all,
>
> we have run into a controversy how to calculate probands' age and
> derive age groups in our epidemiological study.
>
> option 1)
> ------------
> Take the difference in days between day of birth and day of
> examination and divide this by 365.25 (or 365.24) and round the
> result
> to full years, e.g. 45.5000 to 46 and 45.4999 to 45. This will yield
> full year "age estimates" that are nearest to the true value.
>
> option 2) (everyday approach)
> -----------
> Count the number of completed years between day of birth and day of
> examination, as one would do in every day conversations, e.g. if the
> examination is at May 23, 2001 and somebody was born May 23, 1950,
> then he is 51, if the examination was at May 22, 2001, then he is 50
> at examination. This will yield the "every day approach" age in full
> years, but will on average be 0.5 yrs too low compared to the real
> nonrounded value.
>
> If we then compute age ranges like 20 - < 30, we get different
> numbers from these 2 options.
>
> Which one is normally used in publications, when statistics grouped
> bei 5yr or 10yr ranges are to be given?
>
>
> (If requested I will send a summary of replies to the list.)
>
> Thanks in advance for answering.
>
> --
> -----------------------------------------------------------------
> Dietrich Alte (Statistician, Dipl.-Stat.)
> University of Greifswald - Medical Faculty
> Institute of Epidemiology and Social Medicine
> Walther-Rathenau-Str. 48, D-17487 Greifswald, Germany
> Phone +49 (0) 3834 - 86 77 13, fax +49 (0) 3834 - 86 66 84
> Email alte@mail.uni-greifswald.de
> Institute http://www.medizin.uni-greifswald.de/epidem/
> Study http://www.medizin.uni-greifswald.de/epidem/ship.htm
> -----------------------------------------------------------------
=====
---------------------------------------
Dale McLerran
Fred Hutchinson Cancer Research Center
mailto: dmclerra@fhcrc.org
Ph: (206) 667-2926
Fax: (206) 667-5977
---------------------------------------
__________________________________________________
Do You Yahoo!?
Yahoo! Health - your guide to health and wellness
http://health.yahoo.com
| 
