This question was already asked in the forum, but I could not find a clear answer.

It seems that definition of "ClaimsTruncated" changed every year. The "ClaimsTruncated=1" samples have different distribution of number of claims in each year. In the first year number of claims for such samples ranges from 3,5,6 etc. up to ...40,41,42,43. (Number of samples corresponding to last four values are: 237, 217, 116, 581.). In the second year all "ClaimsTruncated=1" samples have same number of claims: 43. And in the third year all "ClaimsTruncated=1" samples have number of claims = 44.

As I understand, the best explanation given or assumed so far in the forum is that the year 3 had cut-off =44; the year 2 had cut-off =43 (why different?), and the year 1 had of course completely different definition, and somehow the difference is related to "ProcedureGroup" data.

Is there any better info on this, any clarification, or definitions of "ClaimsTruncated" that would be easy to understand? Thanks

The truncation was perforrmed on the 95th percentile of the number of claims per member in a year, hence the difference by year. The reason was to make the members with an extreme number of claims not stand out. Those with 43+ (or 44+, depending on the year) were therefore in the top 5% in terms of # claims.

Thanks. That explains difference between cut-off =43 vs =44 in second and third years.

But in the first year among samples with "ClaimsTruncated=1" I see 2 samples with number of claims =3 (not 43 or 44, just 3). There are also 2 samples with number of claims =5. There is 1 sample with number of claims =6. There are 4 samples with number of claims =7. Etc.... goes on through any number of claims - up to 43.

Is there a clear and simple to understand definition on what "ClaimsTruncated=1" meant in teh first year?

Thanks

The ClaimsTruncated field indicates that that member had a large number of claims and some of them were deleted, Claims that were most unique in the population were selected for deletion to meet the 95th percentile cutoff.

"The ClaimsTruncated field indicates that that member had a large number of claims and some of them were deleted, Claims that were most unique in the population were selected for deletion to meet the 95th percentile cutoff."

• This is very strange. This means that many samples of first year had so many unique claims, that only 15 or 10 or 7 or 6 or even just 3 claims were left not deleted per sample. Yet somehow in second year there were no such samples at all: All samples marked as "ClaimsTruncated=1" were left with 43 not deleted claims. In third year all samples marked as "ClaimsTruncated=1" were left with 44 not deleted claims.

Why is so dramatic difference between first year and next two years? Why samples marked as "ClaimsTruncated=1" in the first year had so many "unique claims" that had to be deleted, while in next two years we do not observe this?

Thanks

see this post for an explanation

http://www.heritagehealthprize.com/c/hhp/forums/t/520/claimstruncated/4192#post4192

Thanks, but this is too complicated for me. I do not see a definition of what is "ClaimsTruncated=1" in the first year.

In that thread the answer of arbuckle is not clear to me. And it is followed by two posts with question marks, from Signipinnis and Chris Raimondi.

May be there was a highly logical and clear for everybody definition of "ClaimsTruncated=1" for the first year, may be one can discern it from that the post of arbuckle, but, sorry, it is still a puzzle for me. I do not insist though.

Agree, I didn't really understand the explanation either.