Anyone have any suggestions or thoughts on this?   Much appreciated!     ... View more

Hey sure - I just made up some code to reflect the variables of interest related to figuring out the prevalence:    data Prevalence; input date age hivpos location $ ; datalines; 09/27/19 56 1 A 03/12/15 92 0 B 12/05/16 39 1 D ; I have categorized some of the variables as follows: if year(date) = 2014 then Year = "2014"; else if year(date) = 2015 then Year = "2015"; else if year(date) = 2016 then Year = "2016"; else if year(date) = 2017 then Year = "2017"; else if year(date) = 2018 then Year = "2018"; else if year(date) = 2019 then Year= "2019"; length matage_category $15; if . < age < 20 then age_category = "<20"; else if 20 <= age <= 40 then age_category = "20 to 40"; else if 40 < age <= 60 then age_category = "41 to 60"; else if age > 60 then age_category = ">61"; else if age = . then age_category = "Unknown";   Proc Freq will give me tables with frequencies by age_category / time year / location but divides the tables into HIV=1 and HIV=0 so in principle I could just divide the the cells in each table to get the prevalence however I thought there must be an easy way to internalize this type of calculation. So using the code from the previous post I created a new variable "hivprevalence" which I want to display in tables by the stratified variables above so the output looks something like:  I tried using:  proc means data=prevelence n mean; Class year age_category; var hivprevalence; run; This seemed to get things close - but the hivprevelence calculation is not correct. It dosent seem to be taking the number of hivpos=1 cases and dividing by the total hivpos=0 cases for each cell in the table. Any thoughts / help would be greatly appreciated!       ... View more

Hey @KurtBremser thanks so much! That was super helpful - I was able to get things working with some minor adjustments to my code.    I am now have three new vars:    hivcases hivtotal hivprevalence    I removed the keep statement so that they were added to my existing dataset.    I am having trouble getting the output I intended now, initially I was thinking something like  Proc freq; tables hivprevelence*timeyear; run; which would give me a table with the total prevalence for the disease by year. However, obviously now hivprevelence is not a 1/0 variable and so I used the proc univariate step instead.    proc univariate data=work.import; class timeyear; var hivprevalence; run; This gives me the mean for hivprevelence (which is the prevalence) but it is buried in a mountain of other stuff that is not very useful. I was looking to build an output that is more flexible and table oriented (something more like pivot tables in excel). I tried adding the freq option to proc univariate but that didn't get me too far.    Here is a basic example of the types of tables I am thinking of generating:      Thank you for helping further my learning! Its much appreciated! ... View more