Re: st: Changing values future time points conditional on value from prior time point

[Nick Cox <njcoxstata@gmail.com>]

Note in addition to Jeph's good advice that you have the -if- command
and the -if- qualifier mixed up, presumably due to previous exposure
to some other language. With

if <condition> <action>

there is no tacit loop over all observations. On the contrary, the
command is executed just once. For more, see

http://www.stata.com/support/faqs/programming/if-command-versus-if-qualifier/

Nick
njcoxstata@gmail.com


On 24 February 2014 22:17, Jeph Herrin <info@flyingbuttress.net> wrote:
> this should do
>
>  bys id (time) : replace c_heartcond=1 if c_heartcond[_n-1]==1
>
>
> cheers,
> Jeph
>
>
>
> On 2/24/2014 4:50 PM, Kaufmann, Christopher N. wrote:
>>
>> I am working with data from a longitudinal cohort study with five time
>> points.  The dataset is formatted in a "long" format. I would like to change
>> values of a variable for future time points conditional on the value of the
>> variable in a prior time point. Below is an example of the structure of my
>> dataset in its current form:
>>
>> ID     Time    c_heartcond
>> 1      1       0
>> 1      2       1
>> 1      3       0
>> 1      4       1
>> 1      5       0
>> 2      1       0
>> 2      2       0
>> 2      3       1
>> 2      4       1
>> 2      5       0
>>
>> ID = participant identification number
>> Time = time points of the study
>> c_heartcond = participant reported that they have a heart condition since
>> the last time they were interviewed (1 = yes, 0 = no)
>>
>> As you can see, a person can report a heart condition at one time point,
>> but report that they didn't have a heart condition the next time point. I
>> would like to change the c_heartcond variable so that all future time points
>> after the first that the subject reported the heart condition have a value
>> of 1.  For example:
>>
>> ID     Time    c_heartcond
>> 1      1       0
>> 1      2       1
>> 1      3       1
>> 1      4       1
>> 1      5       1
>> 2      1       0
>> 2      2       0
>> 2      3       1
>> 2      4       1
>> 2      5       1
>>
>> I have written code to do this.  See below:
>>
>> *Fixing comorbidity:
>> foreach var of varlist c_heartcond c_stroke c_hypertension c_pain
>> c_arthritis c_cancer c_diabetes c_incontin c_memorydis {
>>    if c_order_expand == 1 & `var' == 1 {
>>      foreach x1 in 2 3 4 5 {
>>        replace `var' = 1 if c_order_expand == `x1'
>>        }
>>      }
>>    if c_order_expand == 2 & `var' == 1 {
>>      foreach x2 in 3 4 5 {
>>        replace `var' = 1 if c_order_expand == `x2'
>>        }
>>      }
>>    if c_order_expand == 3 & `var' == 1 {
>>      foreach x3 in 4 5 {
>>        replace `var' = 1 if c_order_expand == `x3'
>>        }
>>      }
>>    if c_order_expand == 4 & `var' == 1 {
>>      foreach x4 in 5 {
>>        replace `var' = 1 if c_order_expand == `x4'
>>        }
>>      }
>>    else {
>>      replace `var' = `var' if c_order_expand == 5
>>      }
>>    }
>>
>>
>> But when browsing through the data after running this code, I notice that
>> there are instances where this data is not formatted correctly.  For
>> example, the data might look like this after running my code:
>>
>> ID     Time    c_heartcond
>> 3      1       1
>> 3      2       0
>> 3      3       0
>> 3      4       0
>> 3      5       0
>>
>
*
*   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/