Hi Everyone, I was asked to develop a SAS code the would rescore catergorical variables based on a peer based comparison. I am not event sure how to approach it. I figured a proc freq needs to be used to figure out the count of obs in each category (FMS, SubP, SomeP, WTR, NA, DNR). Then maybe sort obs into they categories. The attache the category count to that observation so I can use it later in an algorithm. Below are the computation direction I have been given. I just really dont event know where to start. and I haven't found anything online that helps answer this question. Thank you to every generous soul who helps!!! Score categorical measures in the following manner: 1. Count the number of hospitals scored in each scoring designations 2. Sum the total number of hospitals across all designations excluding the “N/A” scoring category. 3. For each category, divide the number of hospitals with a scoring designation at or below that category, and divide by the sum from Step 2. 4. Multiply the quotient from Step 3 by 100. 5. Round the result to zero decimal places 6. The results from Step 5 become the scores for each of the scoring designations. Those with the highest possible score designation (“Fully Meets Standard”) receive 100 points, and those designated as “Did Not Respond” get 0 points. Example: A categorical measure has the following hospital counts for each scoring designation FMS =500 SubP =300 SomeP= 200 WTR =100 N/A =150 DNR = 50 In this case, the sum of all hospitals excluding the “N/A” designation is 1150. The score for hospitals with the SubP designation would be the sum of all the hospitals who scored at or below “SubP” (300+200+100+50) divided by 1150. (=650/1150= .5652). Multiplied by 100 and rounded, the score for SubP becomes 57. The same process would be applied in like manner for SomeP and WTR. FMS hospitals would get 100, DNR hospitals would get 0.
... View more