Reputation: 31
and thanks for your help! Without helper columns/rows I'm trying to find the average difference in LVEDi between first and last visits of the patients.
E.g., by hand you would work it out as:
(diff for patient a + diff for pt b + diff for pt c)/3 =
((55-45) + (40-31) + (25-31))/3 =
(10 + 9 + (-6))/3 =
13/3 =
4.333
I'm sure the answer involves indexing with an array function, but I can't get started on how to tackle it!
Any help?
Thanks!
A | B | C | D | E
01 Patient ID | Date | First-Last | LVEDi | LOTS of similar cols
02 a | 3/3/03 | LAST | 55
03 a | 1/1/01 | FIRST | 45
04 a | 2/2/02 | | 65
05 b | | |
06 b | 3/2/16 | LAST | 40
07 b | | |
08 b | | |
09 b | 2/1/12 | FIRST | 31
10 c | 1/7/14 | LAST | 25
11 c | | |
12 c | 2/2/03 | FIRST | 31
13 c | 2/2/08 | | 46
Upvotes: 0
Views: 65
Reputation: 152450
Found a slightly less verbose method:
=SUM(SUMIFS(D:D,C:C,"LAST",A:A,{"a","b","c"})-SUMIFS(D:D,C:C,"FIRST",A:A,{"a","b","c"}))/3
EDIT 1
Building on Lukas' comment:
=(SUMIF(C:C,"LAST",D:D)-SUMIF(C:C,"FIRST",D:D))/COUNTIF(C:C,"LAST")
Upvotes: 1
Reputation: 328
You can use the *
as and AND
criterion in arrays, so with MATCH
you search for the instance, where both conditions multiplied are TRUE
(i.e. 1
). The rest is just filling in your calculation:
{=((INDEX(D2:D13,MATCH(1,(A2:A13="a")*(C2:C13="LAST"),0))-INDEX(D2:D13,MATCH(1,(A2:A13="a")*(C2:C13="FIRST"),0)))+(INDEX(D2:D13,MATCH(1,(A2:A13="b")*(C2:C13="LAST"),0))-INDEX(D2:D13,MATCH(1,(A2:A13="b")*(C2:C13="FIRST"),0)))+(INDEX(D2:D13,MATCH(1,(A2:A13="c")*(C2:C13="LAST"),0))-INDEX(D2:D13,MATCH(1,(A2:A13="c")*(C2:C13="FIRST"),0))))/3}
And remember to use CTRL
+ SHIFT
+ ENTER
to enter the array.
Upvotes: 1