I am working with a medical claims dataset with variables ID (enrollee ID), claimID (a unique identifier for each claim), and the date of service. For this example, assume that claimID 1-5 are all claims for the same medical service, and I only want to count services *within an individual* that occurred during a 7 day period as one service (this is because there are sometimes duplicate claims with dates of service that are in close proximity).     For this example, claimID 1 would be kept and claimID 2 would be deleted because it's within 6 days of claimID 1.  However, claimID 3 would be kept even though it's within 2 days of claimID 2 because claimID 2 will be deleted and claimID 3 is more than 7 days away from claimID 1.  ClaimID 4 and ClaimID 5 would be deleted because they are within 7 days of ClaimID 3, which will be kept.     ID           ClaimID                Date                1                     1                  1/1/2015          1                     2                  1/7/2015          1                     3                  1/9/2015          1                     4                  1/10/2015        1                     5                  1/16/2015          I think the solution might be to create a loop that finds a claim that needs to be eliminated because it's within 7 days of the first claim; as soon as such a claim is detected, the loop exits and the claim is eliminated from the dataset.  On the next iteration of the loop, SAS starts over from the first claim and finds if there is another claim that needs to be eliminated because it's within 7 days; if not, SAS will keep the next the next claim that needs to be deduplicated, etc.  But I'm not sure exactly how to implement this.  I'd appreciate any advice.       ... View more

I have a large claims databases with four diagnosis codes per claim.  The data looks like this:   ID     CLAIM_ID   DATE           DX1        DX2         DX3         DX4 1       100            1/1/2015       7804          1       101            1/1/2015       30921     39021 1       102            2/1/2015       30943     902          01920 1       103            3/1/2015       011 2       104            4/1/2015       4530 2       105            5/1/2015       V9090 3       106            6/1/2015       7039             3       107            6/1/2015       7039 3       108            6/1/2015       E884       0930        1092      0930  3       109            7/1/2015       3094    data have;   infile datalines dlm="," missover;   input id claim_id date $ dx1 $ dx2 $ dx3 $ dx4 $; datalines; 1,100,1/1/2015,7804,,   1,101,1/1/2015,30921,39021,, 1,102,2/1/2015,30943,902,01920, 1,103,3/1/2015,011,,, 2,104,4/1/2015,4530,,, 2,105,5/1/2015,V9090,,, 3,106,6/1/2015,7039,,, 3,107,6/1/2015,7039,,, 3,108,6/1/2015,E884,0930,1092,0930 3,109,7/1/2015,3094,,, ; run;   I want to create a dataset that has all of the *unique* diagnosis codes that occurred on claims during the same day for a given individual, as below (I don't care about whether it was DX1, DX2, DX3, or DX4).  This would involve using proc transpose somehow to create a series of variables that I might call "DXS1-DXS5" (in this fake example, five variables would be created but it would in reality be way more).  I've been playing with the syntax of proc transpose and can't get this work, though.       ID     CLAIM_ID   DATE          DX1     DX2    DX3    DX4   DXS1   DXS2   DXS3   DXS4   DXS5 1       100            1/1/2015       7804                                     7804    30921   39021 1       101            1/1/2015       30921  39021                       7804    30921   39021 1       102            2/1/2015       30943  902     01920            30943  902       01920              1       103            3/1/2015       011                                       011 2       104            4/1/2015       4530                                     4530 2       105            5/1/2015       V9090                                   V9090 3       106            6/1/2015       7039                                     7039   E884     0930    1092    0930 3       107            6/1/2015       7039                                     7039   E884     0930    1092    0930 3       108            6/1/2015       E884    0930  1092   0930   7039   E884     0930    1092    0930  3       109            7/1/2015       3094                                     3094     I'd appreciate suggestions on how to proceed.  Thanks.   ... View more

I have a claims database with enrolle ID, unique claim ID, date, and a field called "Dx_string", which I created by concatatening all of the five-digit ICD-9 diagnosis codes on the claim.  It looks like this:   ID             CLAIM_ID        DATE           DX_STRING 1               100                 1/1/2015       7804          1               101                 1/1/2015       30921 39021 1               102                 2/1/2015       30943 902 01920 1               103                 3/1/2015       011 2               104                 4/1/2015       4530 2               105                 5/1/2015       V9090 3               106                 6/1/2015       E884 3092 3               107                 6/1/2015       7039 3               108                 6/1/2015       800 0930 1092   3               109                 7/1/2015       3094    data have; input ID CLAIM_ID DATE $ DX_STRING $30.; datalines; 1 100 1/1/2015 7804 1 101 1/1/2015 30921 39021 1 102 2/1/2015 30943 902 01920 1 103 3/1/2015 011 2 104 4/1/2015 4530 2 105 5/1/2015 V9090 3 106 6/1/2015 E884 3092 3 107 6/1/2015 7039 3 108 6/1/2015 800 0930 1092 3 109 7/1/2015 3094 ; run;   My goal is to concatenate all DX_STRING values that occur on the same date into a larger string called DX_STRING_DAY:   ID             CLAIM_ID        DATE           DX_STRING            DX_STRING_DAY 1               100                 1/1/2015       7804                        7804 30921 39021 1               101                 1/1/2015       30921 39021           7804 30921 39021 1               102                 2/1/2015       30943 902 01920    30943 902 01920  1               103                 3/1/2015       011                          011 2               104                 4/1/2015       4530                        4530 2               105                 5/1/2015       V9090                      V9090 3               106                 6/1/2015       E884 3092               E844 3092 7039 800 0930 1092 3               107                 6/1/2015       7039                        E844 3092 7039 800 0930 1092  3               108                 6/1/2015       800 0930 1092        E844 3092 7039 800 0930 1092  3               109                 7/1/2015       3094                        3094   I can make the variable called DX_STRING_DAY using this code, but then I can't figure out how to make all values of DX_STRING_DAY the same within an individual on a given day. data want; set have; by id date; retain dx_string_day; if first.date then dx_string_day = dx_string; if not(first.date) then dx_string_day = catx(" ", dx_string_day, dx_string); run;   Any suggestions?  Thanks. ... View more