Re: st: comparing between observations

[Nick Cox <njcoxstata@gmail.com>]

If the focus is on _admission_ dates alone, then -panelthin- (SSC)
sounds right for you.
Nick
njcoxstata@gmail.com


On 24 September 2013 15:46,  <Tina.Shah@uchospitals.edu> wrote:
> Thank you for the quick replies! Regarding Rich's question, if the first observation is the index copd admission, and the second is also for copd but falls within 30 days of the first then the second is not considered an index copd admission.  If the third observation is for COPD and falls more than 30 days from the first, I want to label third observation as a index copd admission also and it does not matter its relation time-wise to the second. Hope that wasn't too convoluted
>
>
> Tina Shah MD MPH
> Pulmonary and Critical Care Fellow
> Pager 3543
>
> -----Original Message-----
> From: owner-statalist@hsphsun2.harvard.edu [mailto:owner-statalist@hsphsun2.harvard.edu] On Behalf Of Richard Goldstein
> Sent: Tuesday, September 24, 2013 8:50 AM
> To: statalist@hsphsun2.harvard.edu
> Subject: Re: st: comparing between observations
>
> note that there is an ambiguity in the question and that Nick's answer makes an assumption about that ambiguity: what if the third observation is not within 30 days of the first discharge but is within 30 days of the second discharge (and the second admission is within 30 days of the first discharge?
>
> Rich
>
> On 9/24/13 9:35 AM, Nick Cox wrote:
>> Check out -panelthin- from SSC. Alternatively, the time since the
>> previous discharge is just
>>
>> gen lapse = admdt - dischdt[_n-1]
>>
>> making sure that you omit all the non-COPD admissions, or equivalently
>> -sort- them aside.
>>
>> You describe your data as if you have just one patient, which seems
>> unlikely, but the previous statement can be generalised by something
>> like
>>
>> bysort patientid (admdt) : gen lapse = ...
>>
>>
>> Nick
>> njcoxstata@gmail.com
>>
>>
>> On 24 September 2013 14:17,  <Tina.Shah@uchospitals.edu> wrote:
>>> I have a large dataset in which I have to identify and index
>>> admission to the
> hospital for copd.  I have created a variable that codes for hospital admission and admission for copd, copdadmit, where 1 equals yes.  The other part of index admission is that the encounter to the hospital cannot be within a 30day window of discharge any other admission for copd.  For example, if patient A has a COPD admission on feb 1to feb 5 and then again is admitted for COPD on feb 15, since feb 15 is within 30days of discharge date of feb 5, this second admission should not be considered an index admission.  I ranked all of the copd admits and I have admission and discharge dates.  How do I identify the copd admits that should not be considered index admits? Also it might be that the second and third ranked observations also fall within 30days of the discharge date from the first.
>>>
>>> My variables and example
>>> Copdadmit, admdt, dischdt, obsrank
>>> 1, 02feb2006, 04feb2006, 1
>>> 1, 22feb2006,  23feb2006, 2
>>> 1, 27feb2006, 28feb2006, 3
>>> 1, 04may2006, 08may2006, 4
>>>
>>> Thanks!
>>> Tina
> *
> *   For searches and help try:
> *   http://www.stata.com/help.cgi?search
> *   http://www.stata.com/support/faqs/resources/statalist-faq/
> *   http://www.ats.ucla.edu/stat/stata/
>
> *
> *   For searches and help try:
> *   http://www.stata.com/help.cgi?search
> *   http://www.stata.com/support/faqs/resources/statalist-faq/
> *   http://www.ats.ucla.edu/stat/stata/

*
*   For searches and help try:
*   http://www.stata.com/help.cgi?search
*   http://www.stata.com/support/faqs/resources/statalist-faq/
*   http://www.ats.ucla.edu/stat/stata/