Recently a fellow coder wrote to me about risk adjusted diagnosis coding. She was responding to an article that I wrote in which I stated the conditions listed in the past medical history should not be included on the claim form by the coder. I stated that in order to include the condition on the claim form, there should be documentation in the history of the present illness or the assessment and the plan that the condition was addressed at the encounter.
She and I were in agreement that we should follow the ICD-10 guideline that states:
“Code all documented conditions that coexist at the time of the encounter and require or affect patient care or treatment.”
Here is her comment:
“However, I am a believer that although the patient is not being seen or treated for the chronic condition or history of condition, they all play a pertinent part in the patient care and overall acuity of the patient.
An example is patient had CVA listed in the PMH and current encounter is for thigh pain without known injury. There could be a correlation to a thrombosis or blood clot that the physician must consider. The physician does not document this correlation, however the old CVA could affect treatment or care. Coding history of CVA code as a secondary would give a clear picture.”
I agree with her: it would give a clearer picture.
Problem lists, and past medical history in electronic health records
But, while it may be relevant to this encounter, without documentation that the clinician was thinking about this past stroke, I would not add it to the claim form. I wouldn’t assume that the physician/NP/PA was thinking about that in terms of the thigh pain after the injury, even though it is on the problem list. In order for me to use the past history of the stroke on the claim form, I would have had to see that the clinician mentioned it in the history of present illness.
For example, the HPI might say,
“he does have a history of stroke, and with his recent injury is at risk of a clot.”
Or, the assessment and plan might say,
“I have reassured him that his symptoms are not consistent with a blood clot, and there is no reason to order an ultrasound.” Or, “I think in light of his history of a stroke, we should do further testing to rule out a thrombolytic event.”
And while I agree with her that the history provides a clearer picture, it is the physician’s job to document a picture.
Hierarchical Condition Categories and MEAT
When I was researching Hierarchical Condition Categories (HCCs) I came across the acronym MEAT.
It stands for:
M–monitoring;
E—evaluating;
A—assessing/ addressing;
T—treatment.
I looked for this in CMS documents and in the ICD-10 guidelines. I didn’t find it specifically, so I suspect that it was a consultant who came up with this acronym. And although I don’t use it in my teaching specifically, I agree with its message. The physician/NP/PA who is providing this service needs to show that a condition “requires or affects patient care or treatment.”
If the condition in question isn’t the presenting problem, the clinician should note that labs were reviewed, history was taken and/or it was considered when developing the assessment and plan. Is not within the scope of my work to make that determination, it is the job of the clinician. If the clinician documented in either the history of present illness or the assessment and plan, then I add that conditioned to the claim form.
Conditions that affect patient care
The example I use the most is the patient who presents to urgent care with a bad case of poison ivy. If the urgent care provider says in the assessment, “I’d like to give her prednisone, but I’m not going to because of her diabetes,” then I add diabetes to the claim form. The diabetes affected patient care and the clinician documented that it affected patient care. If the diabetes is listed in the problem list but not mentioned in the encounter for this date of service, I do not add it to the claim form.
Risk adjustment for history of conditions
But would history of a stroke increase the risk score? That is, if the group has risk based contracts, does adding history of stroke increase the risk score for that patient? The answer is no. The diagnosis codes for current stroke and sequelae of a past stroke (I63, I69) do have HCC weighted scores assigned to him. But past history of the stroke does not. This brings me to a compliance issue in HCC coding.
HCC compliance issues
| Compliance Issues | ||
| Do not use | Use | |
|
Patient seen in office, hospital follow up for stroke |
I63.- Cerebral infarction |
I69.- Sequelae of cerebral infarction |
|
Patient seen in office, ED follow up for TIA |
G45.- Transient cerebral ischemic attacks and related syndromes |
Z86.73 Personal history of transient ischemic attack (TIA), and cerebral infarction without residual deficits |
|
Patient seen and noted to have “history of XXX cancer” but no current evidence of disease or current treatment |
Code indicating malignant neoplasm, starting with the letter C |
Code from category Z85.-, Personal history of malignant neoplasm |
|
Patient seen > 28 days after an acute MI |
I20.-, I21.-, I22.- Current MI |
I25.2 Old myocardial infarction |
Medical groups that are part of Accountable Care Organizations (ACOs) or that have commercial risk based contracts need to assign diagnosis codes carefully.
Reviewed, Betsy Nicoletti 10/12/2018
Additional resources
The CodingIntel Guide to Hierarchical Condition Categories provides a comprehensive list of HCC and Risk Adjusted Diagnosis Coding resources available on CodingIntel.
