I am analysing hospital admission data to find alcohol-related cases and assign a fraction, indicating the proportion of cases that are alcohol-caused.

The analysis involves creating a variable called alc indicating that the admission is alcohol-related. This is based on the diagnostic code and there are 45 values of alc. Then a fraction (a new variable called af) is applied, based on the value of alc, and the age group and sex of the patient.

I have done similar analysis in the past, but only on the principal diagnosis. This time I need to include the 13 secondary diagnoses and use only the highest fraction.

Can anyone tell me how I can do the following in SPSS:

for each diagnosis (variables diag1 to diag14), assign the alc values (there will need to be 14 alc variables, one for each diagnosis).
For each alc variable assign a fraction (af1 to af14) based on the value of alc, and the age group and sex of the patient.
Check which of the af variables has the highest value and put this value into another variable (maxaf).

I hope this is clear, if not I can post the program I've written to get it working for one diagnosis if that helps.

I'm on a tight time frame with this analysis so any help people can give me to save time wasted in trial and error will be much appreciated.

Thanks in anticipation!

The alc values and fractions are related to the other variables using some math formula (like af1+age-10*sex), right?
You can compute new variables for both:


For the maxaf variable, use COMPUTE again, but with tha MAX function:



It would be good if you could post your commands for this, maybe you're thinking in a faster way and maybe someone can point any common mistakes.

Dear Fierce,

Thanks for your reply! The compute max command is just what I was looking for.

However, the fractions are not calculated using a maths formula, they have been assigned by meta-analysis of the scientific literature. I'll give you an example of the coding I've done so you can see what I mean.

There are two programs, the first creates the alc variable based on the diagnostic code. There are 45 conditions that are alcohol-related so I'll just include a few.

IF (diag1 >= 'C15 ' and diag1 <= 'C159') alc1= 15.
IF (diag1 >= 'C18 ' and diag1 <= 'C189') alc1= 16.
IF (diag1 >= 'C20 ' and diag1 <= 'C209') alc1= 17.

The next program assigns the fraction, according to the value of alc, agegroup and sex. The example below is for males only.

DO IF (alc1 = 15 & sex = 1) .
RECODE
agegp
(2=0.32) (3=0.31) (4=0.30) (5=0.34) (6=0.32) (7=0.26) (8=0.20) INTO AF1 .
END IF .

DO IF (alc1 = 16 & sex = 1) .
RECODE
agegp
(2=0.05) (3=0.05) (4=0.04) (5=0.05) (6=0.05) (7=0.04) (8=0.03) INTO AF1 .
END IF .

DO IF (alc1 = 17 & sex = 1) .
RECODE
agegp
(2 thru 4=0.08) (5=0.09) (6=0.08) (7=0.07) (8=0.05) INTO AF1 .
END IF .

Given there are 45 conditions and 14 diagnosic fields, the programs end up huge if I paste and copy each program 14 times to process each diagnostic field. I was wondering if there's a command I can use at the beginning of each program to say 'do the following for diag 1 to 14 / alc 1 to 14.'

Hope this has made it clearer. It's very difficult to explain programming in English!

Kind regards,

Liz

Hey guys,

I think I’ve solved how to do the iterative analysis I posted a question about a couple of days ago. Just thought I’d share the solution in case anyone else has the same problem.

For the first program the syntax is:

DO REPEAT diag = diag1 TO diag14
/alc = alc1 TO alc14.

[put the IF statements here]

END REPEAT.

For the second program I just replaced diag with alc, and alc with af in the syntax above.


On another topic, if anyone reading this can tell me why the following code doesn’t remove trailing Xs from my data file, I’d be very grateful. I posted this query in the Enhancing Data section earlier. The syntax works for dashes but not Xs. Weird!

string newerdiag (A6).
compute newerdiag = RTRIM(newdiag, 'X').
execute.

Thanks.

Liz