Monday, April 02, 2012

Divide by Zero: Weird Math in CMS Clinical Quality Measure (CQM) Criteria

From the CMS "Medicare Electronic Health Record (EHR) Incentive Program - ATTESTATION USER GUIDE For Eligible Professionals (EPs)" (warning: large PDF), page 41/64:

... Step 25 – Core Clinical Quality Measures (CQMs 1 of 3)

EPs must report calculated CQMs directly from their certified EHR technology as a requirement of the EHR Incentive Programs. Each EP must report on three core CQMs (or alternate core) and three additional quality measures. If one or more core CQMs is outside your scope of practice, you will have to report on an equal number of alternate core CQM(s).

If the denominator value for all three of the core CQMs is zero, an EP must report a zero denominator for all such core measures, and then must also report on all three alternate core CQMs.

If the denominator value for all three of the alternate core CQMs is also zero an EP still needs to report on three additional clinical quality measures. Zero is an acceptable denominator provided that this value was produced by certified EHR technology.

Now, while I had an "800" in the math section of the SAT, where I believed that a fraction with a denominator of zero had a value of either infinity or 'undefined', that was many moons ago. Perhaps my knowledge of mathematics is now obsolete...

Wait - I tried this simple program on an old Microsoft MS-DOS GWBASIC interpreter, ported ca. 1981 to the Intel 8086/88 from Bill Gates' original 8080-based MBASIC, that I have laying around on my PC:

10 INPUT X
20 LET Y=X/0
30 PRINT Y
40 END


and got this warning/error message, right from Bill Gates:

"Division by zero"

and the answer: 1.701412E+38 (infinity in the 8/16-bit world from where GWBASIC sprang).

So ... allow me to say I find CMS math just a bit puzzling.

Wait ... now I understand.

Infinite quality! :-)

But thank heavens the zero denominator is only accepted when produced by 'certified' health IT.

Uncertified health IT is liable to produce a denominator of "i" (that is, the square root of -1).

-- SS

Addendum: since I am not a government math genius, I checked with Wikipedia:

... A common fraction (also known as a vulgar fraction or simple fraction) is a rational number written as a/b or \tfrac{a}{b}, where the integers a and b are called the numerator and the denominator, respectively.[1] The numerator represents a number of equal parts and the denominator, which cannot be zero, indicates how many of those parts make up a unit or a whole.


-- SS

Addendum:

This post is partly satire. I was a day late for April 1 but...

-- SS

7 comments:

RT PCR Primers said...

Hi,
They probably believe that with cut and paste technology from interactive EHRs, they can upcode their way out of the red....!!

Exceptionally Brash said...

Wow! Are you sure you didn't miss this post by a day?

RoseAG said...

In every program I've written if the denominator is zero I don't do the division. It's OK for the denominator to be zero, the program logic is set to skip around it.

Most likely they want a zero reported rather than a missing value. A missing value raises the question as to whether the value is zero or is something else but just wasn't reported correctly.

InformaticsMD said...

Re: RoseAG said...

the program logic is set to skip around it.

This post is meant to be, in part, tongue in cheek.

-- SS

InformaticsMD said...

Exceptionally Brash said...

Wow! Are you sure you didn't miss this post by a day?

If you left a link, it did not come through.

-- SS

Live IT or live with IT said...

Oh you are so negative! What solution do you offer for the divide by zero problem?

InformaticsMD said...

Live IT or live with IT said...

Oh you are so negative! What solution do you offer for the divide by zero problem?

Forced use of hexadecimal only by the perps in all their future bank statements.