Dr Simon Clay continues his series looking at the importance of episode coding for patients with CKD and why some patients can be erroneously missed off the register.
This article will go through recent changes to QOF coding for chronic kidney disease, which can be tricky. This article is correct up to the changes made in December 2009 (v 16.0).
Proteinuria definitions
OUT (codes that are no longer valid)
• 4674 (urine protein test = +)
• 4675 (urine protein test = ++)
• 4676 (urine protein test = +++)
• 4677 (urine protein test = ++++)
IN: (new valid codes)
• 1Z1B. (CKD stage 3 with proteinuria)
• 1Z1D. (CKD stage 3a with proteinuria)
• 1Z1F. (CKD stage 3b with proteinuria)
• 1Z1H. (CKD stage 4 with proteinuria)
• 1Z1K. (CKD stage 5 with proteinuria)
The following proteinuria codes remain valid:
• R110 ([D]Proteinuria)
• R1100 ([D]Albuminuria)
• R1103 ([D]Microalbuminuria)
• R110z ([D]Proteinuria NOS)
No time criteria are applied to any of the proteinuria codes. Any presence in the record will count.
A new indicator, CKD6
CKD6 requires that all patients with CKD stage 3 or worse need an albumin-creatinine or protein-creatinine ratio test annually. This obligation remains, even if the patient has documented proteinuria or microalbuminuria already.
To record ACR or PCR (ACR is much preferred in renal circles and laboratories are supposed to be sorting out the logistics to offer this test, I understand) the codes are as follows:
• ACR (46TC)
• PCR (44lD) (that is lower case ‘L' then D)
These need to be recorded every 15 months – giving practices the same latitude as usual about codes being added in January to March (therefore counting for two consecutive years).
Severity divisions remain, though they are not enforced by the rules so subdivision into stages 3a and 3b is not required (CKD3a is defined as eGFR of 45-59ml/min and CKD3b as eGFR of 30-44 ml/min).
Definition of proteinuria
What constitutes a positive test for proteinuria? According to NICE guidance, an ACR of 30 should be regarded as positive in patients without diabetes and 2.5 or 3.5 in male and female patients with diabetes, respectively.
Testing for proteinuria only affects CKD5 (the requirement to start ACE inhibitor treatment). If a patient's urine is never checked for protein, that patient is never entered into the denominator for CKD5.
This is good news financially but not good clinically as any software a practice is using to help remind them to consider ACE's will not give such a prompt in hypertensive CKD patients unless they have also been coded as proteinuric.
Having listed some of the valid CKD codes above, but only those which also define the patient as being proteinuric, I just want to avoid confusion by listing all the valid CKD codes (including the ones which suggest the patient is not proteinuric). These are:
• 1Z12 (chronic kidney disease stage 3)
• 1Z13 (chronic kidney disease stage 4)
• 1Z14 (chronic kidney disease stage 5)
• 1Z15 (chronic kidney disease stage 3A - v12)
• 1Z16 (chronic kidney disease stage 3B - v12)
• 1Z1B (chronic kidney disease stage 3 with proteinuria - v12)
• 1Z1C (chronic kidney disease stage 3 without proteinuria - v12)
• 1Z1D (chronic kidney disease stage 3A with proteinuria - v12)
• 1Z1E (chronic kidney disease stage 3A without proteinuria - v12)
• 1Z1F (chronic kidney disease stage 3B with proteinuria - v12)
• 1Z1G (chronic kidney disease stage 3B without proteinuria - v12)
• 1Z1H (chronic kidney disease stage 4 with proteinuria - v12)
• 1Z1J (chronic kidney disease stage 4 without proteinuria - v12)
• 1Z1K (chronic kidney disease stage 5 with proteinuria - v12)
• 1Z1L (chronic kidney disease stage 5 without proteinuria - v12)
Episode coding
The issue of episode codes should finally be mentioned. Each CKD code should be episode coded, but many practices do not bother.
The rulesets require that the analysis software looks for the presence of a CKD3 or worse code. The presence of a CKD2 or better code should also be looked for. These two concepts are treated separately and the last of each type of code which is episode-coded as ‘first' or ‘new' is looked for and both are seen by the software. The software then applies the rules' logic to determine the patient's correct CKD grade.
In the case of CKD codes of CKD2 or better, the ruleset defines valid codes as: ‘Latest first or new episode < (REF_DAT) AND > date of diagnostic code above' (where the ‘diagnostic code above' is the list of CKD codes of grade 3 or worse).
The final determination of the patient's grade does depend on whether or not the practice is routinely adding an episode code of ‘new' to each new CKD code as it's entered.
Here are some examples of how the rules are applied in the real world:
1. All codes episode coded:
a.) CKD2 code, followed later by CKD3 code: Both codes are seen by the software and the later CKD3 code is correctly seen as the valid code. Patient is included on the CKD register.
b.) CKD3 code followed by CKD2 code: Both codes are seen by the software and the later CKD2 code is correctly seen as the valid code. Patient is excluded from the CKD register.
c.) CKD2 code, followed by CKD3 code, followed by CKD2 code: Software notes CKD3 code, but also notes that there is a later CKD2 code which fulfils the definition above. Later CKD2 code is therefore seen as the valid code. Patient is excluded from the CKD register.
d.) CKD3 code, followed by CKD2 code, followed by CKD3 code: Software notes presence of CKD2 code, but also that there is a later CKD3 code. This latter code is recognised as the valid code and the patient is included in the CKD register.
2. No codes episode coded:
a.) CKD2 code, followed later by CKD3 code: Both codes seen by the software, CKD2 code does not fulfil the requirement to be after the CKD3 code, hence is not valid. This leaves the CKD3 codes to be the valid code – correctly. Patient is correctly included in the CKD register.
b.) CKD3 code followed by CKD2 code: Both codes are seen by the software, despite the lack of episode coding, because the two families of codes are looked for separately, the later CKD2 code is correctly seen as the valid code. Patient is correctly excluded from the CKD register.
c.) CKD2 code, followed by CKD3 code, followed by CKD2 code: Software notes the first CKD2 code and the CKD3 code, but due to the absence of episode coding, the second CKD2 code is ignored in favour of the first one. Since the first CKD2 code does not come after any CKD3 or worse CKD code, the CKD3 code is seen as the valid code and the patient is erroneously included in the CKD register.
d.) CKD3 code, followed by CKD2 code, followed by CKD3 code: Software notes presence of CKD2 code, but the later CKD3 code is ignored in favour of the first ever CKD3 or worse code (due to the absence of episode coding). The software perceives the CKD2 code as the valid code and the patient is erroneously excluded from the CKD register.
So in the absence of episode coding, there are a number of scenarios where the software may incorrectly interpret the situation of a patient whose CKD grade is see-sawing between CKD2 and CKD3 or worse, and they may be erroneously included or excluded from the CKD register.
Dr Simon Clay is a GP in Erdington, Birmingham
QOF coding tips: Chronic kidney disease Urine testingClick here for more advice on issues relating to urine testing for proteinura
