The Problem List Is Not the Assessment: Diabetes and Kidney Disease in a Three-Problem Visit
Mr. Alvarez is a 68-year-old man in a Medicare Advantage plan who is seen for a routine follow-up at his primary care office on March 12. The provider's assessment reads: 'Type 2 diabetes mellitus, stable - continue metformin (held today). Hypertension, controlled. Renal insufficiency, stable, nephrology following.' The lab section of the same note lists eGFR 32 mL/min/1.73 m², creatinine 1.9 mg/dL, and hemoglobin A1c 7.1%. The clinic's problem list, maintained by front-desk staff and not part of the provider's assessment, shows 'DM with CKD stage 3,' 'History of DKA 2019,' and 'HTN.' Nothing in the provider's own words links the diabetes to the kidney disease, and the CKD stage appears only on the problem list. The plan's encounter data submission deadline is in three days and the provider has left for a two-week vacation.
1. Which diagnoses from this encounter can you submit today, and which cannot? Walk through each item in the record and explain your reasoning.
Read the worked answer
Submittable now
- I10, essential hypertension. Documented by the provider in the assessment; valid, though it carries no HCC.
- E11.9, type 2 diabetes mellitus without complications. The provider wrote 'type 2 diabetes mellitus' and said nothing about complications. Coding diabetes as 'with complications' requires the provider to describe a complication; the coder cannot supply it from labs or a problem list.
Not submittable as documented
- A diabetes-with-CKD combination (E11.22 plus N18.3-). This requires (a) a provider statement linking the two and (b) a documented stage. The stage exists only on a staff-maintained problem list, which is not encounter documentation, and the eGFR of 32 is clinical data, not a provider diagnosis. Assigning N18.3 from a lab value is inference, and inference is exactly what RADV reviewers look for.
- 'Renal insufficiency.' The term is not a stage-specific CKD code and should not be assumed to mean CKD stage 3. Even coded generically it would carry no HCC.
- 'History of DKA 2019.' Historical, not a current condition, and not HCC-bearing in any event.
The correct move is to keep E11.9 and I10, open a compliant query, and hold the diabetes-CKD codes. If the reply arrives after the deadline, submit the corrected diagnosis in the next data cycle rather than guessing now.
2. Write the query you would send, and explain why each element is designed the way it is.
Read the worked answer
Draft query (non-leading, multiple choice, with clinical indicators cited):
'Mr. Alvarez's note of March 12 documents type 2 diabetes mellitus and renal insufficiency. The eGFR is 32 mL/min/1.73 m² and the problem list references CKD stage 3. Please clarify:
- Is the chronic kidney disease related to the diabetes (diabetic kidney disease), or is it a separate/other cause?
- If CKD is present, what stage is documented - stage 3a, 3b, 4, 5, or ESRD?
- If diabetic kidney disease is present, please add the linkage and stage to your assessment. Reply: ______ / Unable to determine: ______'
Why it is built this way
- It offers options rather than proposing the answer; a query that says 'would you agree this is diabetic CKD?' is leading and is itself an audit finding.
- It quotes the actual supporting data (eGFR, problem list) so the provider is responding to evidence in the record, not to the coder's hunch.
- It asks two separate questions, because linkage and stage are separate documentation requirements and a single 'yes' would not support a code.
- It contains no diagnosis the provider has not already mentioned, so it cannot be read as prompting a new condition.
- An 'unable to determine' option keeps the record honest; without it, providers often pick a choice just to clear the queue.