Determine interval with at least 7 non-missing values

[Sasu1]

Hey everyone,

I have to do some analysis on a pharmaceutical dataset.

By selecting the right interval I lack the right idea for continuing.

I would be grateful if someone could help me 🙂

My data:

I am working with a dataset which contains measurements (like temperature, weight etc.) from patients over different timepoints.

Patient      Day          Parameter              Value

1                 1              Temperature            37

                   2                                              38.5

                   3                                              39

                   4                                               -

                   5                                               -

                   6                                               35.9

                   7                                               38

                   8                                               -

                   9                                               -

                  10                                              38.5

                  11                                               -

                  12                                              39.3

                  13                                              -

                  14                                              38

                  15                                               -

                  16                                              36.9

                  17                                              -

2                 1                                                -

                  ...                                               ...

                   23                                              37.6

My task:

For my analysis I want to determine a 14-days interval in which at least 7 values are not missing.

I am starting with the measurement on the last day (so first I sort the dataset in descending sequence). therefore for patient 1 with the intervall [Day 17 to Day 4]. In this example the interval doesn't contain 7 non-missing values.
So I have to move the interval from [Day 17 to 4] to [Day 16 to 3] and have to check the number of non-missing values
again. Now there are 7 non-missing values in the interval, I can give out the start and end day of the interval and use it for analysis.
So all in all, I start with the interval [last Day - (last Day - 14)] and move it forward until I reach an interval which contains at
least 7 non-missing values. This interval should be determined for each patient and each parameter.
             

My solution:

First I sort the dataset in descending sequence. After that I have transpose the dataset twice - for the variable Day and Value. Then I have created two arrays - _days and _values.
Now I count the missing values in the interval from position 1 to 14. If the value is greater than 7, then I have to move the interval. I have tried to realize it with a do-while loop.

Does everyone has an idea how to solve that problem?

Thanks for your help!

Kind regards,

Saskia

[mkeintz]

What do you want the output to look like?

[s_lassen]

I think that you want a zero-based array of length 14, something like this:

data want;
  array measures(0:13) 8 _temporary_;
  call missing(of measures(*));
  do _N_=1 by 1 until(last.patient);  
    set have;
    by patient;
    measures(mod(_N_,14))=parameter;
    if _N_>=14 and n(of measures(*))>=7 then
      output;
   end;
run;

That way, the array will just keep cycling, and will always contain the latest 14 values.