Solved
New Contributor
Posts: 2

# creating a new variable based on a frequency

Hello,

I am dealing with data where I have subject IDs and a variable that tells me the diseases each subject has. Some of the subjects have more than one disease but they are just inputted in the dataset multiple times. For example:

data example;

input ID condition\$ ;

datalines;

45 diabetes

46 cancer

46 diabetes

47 cancer

What I want is a summary of the frequency of conditions. For example, X number of people have only one condition but Y number of people have two conditions and Z number of people have three conditions, etc. I am thinking that I would just like to create a variable that tells me the number of conditions each participant has but I am not sure how to do this.

Is there a way I can do this?

Accepted Solutions
Solution
‎10-01-2017 02:39 PM
Regular Contributor
Posts: 213

## Re: creating a new variable based on a frequency

Use proc summary to get the number of conditions each patient has, then pro freq to get the required output.

``````proc summary data=example nway;
class id;
output out=CondById(drop=_type_ rename=(_freq_=NumConditions));
run;

proc freq data=CondById noprint;
tables NumConditions / out=FreqCond(drop=percent);
run;``````

All Replies
Super User
Posts: 23,776

## Re: creating a new variable based on a frequency

Is a patient going to have multiple diagnosis entries for the same event, ie will diabetes be there twice?

If not, you can use a PROC FREQ to get the number of counts. Otherwise, you're looking to count distinct.

I just posted a solution on how to count distinct items per group here:

https://communities.sas.com/t5/SAS-Statistical-Procedures/Count-unique-variables-associated-with-one...

Solution
‎10-01-2017 02:39 PM
Regular Contributor
Posts: 213

## Re: creating a new variable based on a frequency

Use proc summary to get the number of conditions each patient has, then pro freq to get the required output.

``````proc summary data=example nway;
class id;
output out=CondById(drop=_type_ rename=(_freq_=NumConditions));
run;

proc freq data=CondById noprint;
tables NumConditions / out=FreqCond(drop=percent);
run;``````

New Contributor
Posts: 2