Re: st: Need Help with Stata Programming

[Nick Cox <njcoxstata@gmail.com>]

In that case, date of event for each patient is


egen date_of_event =  total(event * date), by(patientid)

or

egen date_of_event = mean(date / event), by(patientid)

There are other ways to get it too, but two is more than you need. You
may need to spread the information to all observations for each
patient.
Nick
njcoxstata@gmail.com


On 21 November 2013 15:04, Michael Stewart
<michaelstewartresearch@gmail.com> wrote:
> Hello Nick.
>
> I apologize that I missed your question.
> Yes it is true that the event can only happen once for a patient.
> Thanks , always for your help and useful articles/references.
>
> On Thu, Nov 21, 2013 at 9:59 AM, Nick Cox <njcoxstata@gmail.com> wrote:
>> What you report as the difficulty is explained in detail in the paper
>> already cited
>>
>> SJ-7-3 pr0033 . . . . . . . . . . . . . . Stata tip 51: Events in intervals
>> . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . N. J. Cox
>> Q3/07 SJ 7(3):440--443 (no commands)
>> tip for counting or summarizing irregularly spaced
>> events in intervals
>>
>> which is accessible at http://www.stata-journal.com/sjpdf.html?articlenum=pr0033
>>
>> In short, you need an condition on the dates that enter in the
>> calculation. It is convenient, but not essential, to use -inrange()-,
>> e.g.
>>
>> ... inrange(date, <date of event> + 1, <date of event> + 60)
>>
>> ... inrange(date, <date of event> - 60, <date of event> - 1)
>>
>> where naturally you need to fill in <date of event>.
>>
>> You haven't answered my question about whether the event occurs at
>> most once for each patient. Regardless, the date of the last event can
>> always be found by
>>
>> gen date_of_event = date if event
>>
>> bysort patientid (date) : replace date_of_event = date_of_event[_n-1]
>> if missing(date_of_event)
>>
>> and the date of the next event can be calculated using the time
>> reversal method discussed in
>>
>> http://www.stata.com/support/faqs/data-management/replacing-missing-values/
>>
>> Nick
>> njcoxstata@gmail.com
>>
>>
>> On 21 November 2013 03:38, Michael Stewart
>> <michaelstewartresearch@gmail.com> wrote:
>>> Dear Nick ,
>>>
>>> Thank you very much for the advice .I looked at the prior code, your
>>> reference(took a print out) and read it and tried to solve the problem
>>> at hand.
>>>
>>>
>>> My Goal :
>>> 1)Count number of admissions(i.e., admit_type==IP) 60 days before the
>>> event(i.e., event==1) OR 60 days after the event(i.e., event==1).
>>> 2)Sum up the cost of the admissions (i.e., admit_type==IP) 60 days
>>> before the event(i.e., event==1) OR 60 days after the event(i.e.,
>>> event==1).
>>>
>>> I am trying to solve the problem with hints amd have been stuck up here
>>>
>>> ***************************************************************************
>>> gen NUM_IP_PRIOR=.
>>>
>>> qui forvalues i=1/`=_N'{
>>> sum admit_type if admit_type==IP & Pt_id==Pt_id[`i']
>>> replace NUM_IP_PRIOR=r(N) in `i'
>>> }
>>> *************************************************************************
>>>
>>> I am able to get number of IP (inpatient admissions) for each patient
>>> as a whole but not able to further break it down by
>>> (i)before and after the event
>>> (ii)60 days before the event(i.e., event==1) OR 60 days after the
>>> event(i.e., event==1).[with a time component]
>>>
>>> How could I add the condition/stata pseudocode so as to get count of
>>> IP before the event(variable-1) OR after the event(variable-2)
>>> .Please help.
>>>
>>>
>>> Thank you very much for your time
>>>
>>> On Wed, Nov 20, 2013 at 8:27 AM, Nick Cox <njcoxstata@gmail.com> wrote:
>>>> A similar question was posted only yesterday
>>>>
>>>> http://www.stata.com/statalist/archive/2013-11/msg00745.html
>>>>
>>>> and advice included a reference
>>>>
>>>> SJ-7-3  pr0033  . . . . . . . . . . . . . .  Stata tip 51: Events in intervals
>>>>         . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  N. J. Cox
>>>>         Q3/07   SJ 7(3):440--443                                 (no commands)
>>>>         tip for counting or summarizing irregularly spaced
>>>>         events in intervals
>>>>
>>>> which is accessible at http://www.stata-journal.com/sjpdf.html?articlenum=pr0033
>>>>
>>>>
>>>> Nick
>>>> njcoxstata@gmail.com
>>>>
>>>>
>>>> On 20 November 2013 13:11, Michael Stewart
>>>> <michaelstewartresearch@gmail.com> wrote:
>>>>
>>>>> I am trying to measure number of visits/admissions before and after a
>>>>> particular event and was hoping to get help from Stata programmers in
>>>>> the Statalist
>>>>>
>>>>> My dataset is as follows . ED: stands for emergency department visit &
>>>>> IP stands for Inpatient admission
>>>>>
>>>>> Patients are sorted chronologically (by date).
>>>>>
>>>>> My goal is
>>>>>
>>>>> 1)Count number of admissions(i.e., admit_type==IP) 60 days before the
>>>>> event(i.e., event==1) OR 60 days after the event(i.e., event==1).
>>>>> 2)Sum up the cost of the admissions (i.e., admit_type==IP) 60 days
>>>>> before the event(i.e., event==1) OR 60 days after the event(i.e.,
>>>>> event==1).
>>>>
>>>>
>>>>> 3               ED              0               date_1  USD xx
>>>>> 3               IP              0               date_2  USD xx
>>>>> 3               ED              0               date_3  USD xx
>>>>> 3               IP              0               date_4  USD xx
>>>>> 3               IP              1               date_5  USD xx
>>>>> 3               ED              0               date_6  USD xx
>>>>> 3               IP              0               date_7  USD xx
>>>>> 3               ED              0               date_8  USD xx
>>>> *
>>>> *   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/
>>>
>>>
>>>
>>> --
>>> Thank you ,
>>> Yours Sincerely,
>>> Mike.
>>> *
>>> *   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/
>
>
>
> --
> Thank you ,
> Yours Sincerely,
> Mike.
> *
> *   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/