Medical Coding for Risk Adjustment
While all risk adjustment payment models differ in some areas, one common ground they hold is diagnosis coding. Risk adjustment is a payment methodology that uses ICD-10-CM codes, organized into Hierarchical Condition Categories (HCCs), to establish a risk score for each patient.
Medical coders have a special role when it comes to coding for risk adjustment, and there are measures they can implement to increase coding accuracy. Whether an HCC coder (also referred to as a risk adjustment coder) is coding for a physician’s office, a health plan, or a government auditing contractor, they need to understand the complexity of diseases associated with chronic conditions or comorbidities to ensure the documentation supports the accurate health status of the patient. Medical terminology, anatomy, and pharmacology are additional areas of expertise required of competent risk adjustment coders.
Keep in mind, what might be “good enough” to establish medical necessity on the fee-for-service (FFS) claim may not be specific enough for accurate risk score calculation. As Scenario 1 shows, HCC coding relies on all documentation available, not just the provider’s final assessment, for a date of service. The scenarios below use HCCs from the Centers for Medicare & Medicaid Services (CMS).
A provider documented in the medical record details about their 65-year-old patient who recently enrolled in an MAO. She came in for her Welcome to Medicare office visit. Of special note, the medical assistant documented a depression screen result of 11 which indicated possible moderate depression. The patient had a diagnosis of depression in the problem list and had been on an antidepressant for about six months with no notable improvement. The physician addressed the questionnaire with the patient, asked some more pertinent questions, and listed moderate recurrent major depression in the final assessment.
The coder entered the correct service code along with the diagnosis codes of Z00.01 Encounter for general adult medical examination with abnormal findings and F32.A Depression, unspecified.
The claim was paid; however, during a routine audit as outlined in the office’s clinical documentation improvement plan, an auditor noticed that the coder did not accurately code the depression as diagnosed by the provider (F33.1 Major depressive disorder, recurrent, moderate). While this particular change in diagnosis coding would not affect the payer’s decision about medical necessity and payment, it does affect the risk score calculation of the patient.
F32.A = no mapping to CMS-HCC/no risk score
F33.1 = CMS-HCC 59/score value of 0.309
As much as it is important to accurately capture all conditions that currently exist and require treatment, it is equally important to not submit diagnosis codes for conditions the documentation does not support. For instance, when the full documentation for the encounter provides more accurate information for coding purposes than the assessment does, the coder should base code choice on the full documentation, as shown in Scenario 2.
A 72-year-old male patient presented to the office for a routine follow-up of ongoing residual left-sided weakness due to his stroke last year. A detailed exam was performed. The patient stated he was feeling well, was taking his medications as prescribed, and had help at home to aid with his limited walking stability and other Activities of Daily Living. He refused a wheelchair or physical therapy at this time. The provider listed stroke in the final assessment.
The coder reviewed the documentation of the office visit before submitting the claim and accurately coded I69.354 Hemiplegia and hemiparesis following cerebral infarction affecting left non-dominant side.
The risk adjustment coder knew of AHA Coding Clinic® for ICD-10-CM and ICD-10-PCS guidance from 2015, which instructs coders to code weakness due to stroke as hemiparesis. The coder also applied ICD-10-CM Official Guidelines for Coding and Reporting, Section I.C.9.d, concerning coding sequelae of cerebrovascular disease to capture the late effect of the stroke instead of coding an acute cerebrovascular infarction (I63.9 Cerebral infarction, unspecified) after the acute phase of the stroke has resolved.
I63.9 = CMS-HCC 100/score value of 0.230
I69.354 = CMS-HCC 103/score value of 0.437
CMS has made it clear that it is the responsibility of the health plan to not only gather diagnosis codes that support specific HCCs, but to also look for overcoded conditions. Overcoding in risk adjustment refers to using an incorrect code with a higher score value rather than the correct code based on documentation. Scenario 3 provides an example of how to prevent overcoding.
A 67-year-old female presented to the office for follow-up of an ulcer on her left calf. Previously the provider had documented this was related to her atherosclerosis and ordered wound care. In today’s exam the provider noted the ulcer was healed, but in the assessment the provider continued to document atherosclerosis with left calf ulcer. The electronic medical record (EMR) system chose I70.242 Atherosclerosis of native arteries of left leg with ulceration of calf for this diagnosis.
The coder reviewed the inconsistent documentation and queried the provider. The correct code of I70.202 Unspecified atherosclerosis of native arteries of extremities, left leg was submitted.
I70.242 = CMS-HCC 106/score value of 1.488
I70.202 = CMS-HCC 108/score value of 0.288
There may be instances in which the coder cannot make an educated determination of the correct code and clarification from the provider is necessary prior to claim submission. If warranted, the provider may attach an addendum to the office note clarifying the documentation.
Most likely though, a coder’s communication will come in the form of documentation improvement training, as the next section describes, based on findings during a retrospective review and not from a concurrent query.