With paying for value ascendant, diagnostic coding accuracy grows more crucial

Jennifer SwindleCoding

Value-based payment has been steadily growing in healthcare in the past few years, thanks to programs such as the Hospital Inpatient Value-based Payment Program and the Merit-based Incentive Payment System (MIPS, part of the Quality Payment Program for physicians). More recently, this movement has gained new agency due to risk-based programs in the commercial payer market. For providers, these initiatives, while promoting innovation and safety in clinical care, also put diagnosis coding under the microscope.

Coding at the highest level of accuracy and specificity has always been instrumental in supporting claims of medical necessity and painting a picture of the clinical condition of the patient. Too often, however, providers have bumped along with incomplete diagnosis coding, leaving a bare sketch of the patient’s condition, not truly a finished picture. With value-based payment apparently here to stay – and perhaps to become a much bigger amount of reimbursement at risk – providers that want to stay in business must be exacting in diagnostic coding accuracy.

Outpatient providers have learned this lesson faster, because the regulations are clearer in regard to capturing definitive diagnoses. They already know that documentation with phrases such as “consistent with,” “compatible to,” “possibly,” etc., likely will be kicked back to them with a denied claim.

Providers face many areas of risk in documentation; here are just a few:

  • Every encounter, regardless of setting, needs to have a clearly stated or easily inferred chief complaint. Even on an inpatient facility record, where a provider or possibly multiple providers see the same patient daily, each individual provider who sees that patient must document a chief complaint every day to show that what he or she was doing was medically necessary.
  • When you open a patient’s electronic health record, some documentation is auto filled, so you need to go in and edit those fields for the current visit.
  • Voice recognition/natural language processing systems may misinterpret something, causing documentation to be incorrect. Careful proofreading and editing is required by the provider.Problem lists are often not well maintained and providers hesitate to remove things they did not add; however, if the problem list pulls into a current note and is left unedited and the note is authenticated, that provider is indicating that problem list is true and accurate on that date.

Patients often present, particularly in the office or emergency department setting, with one complaint, but at the end of the visit, there are multiple conditions listed. These may be problems the patient has had historically; however, if there is no evidence they were addressed or impacted the care, a list does not support coding those conditions.

Providers are the only individuals who can diagnose a patient, so elements in the chart that are clinically abnormal cannot be coded until the provider makes the diagnosis. Abnormally low pulse oximetry, abnormal lab values and abnormal echocardiogram tracings may be evident to the provider of a condition, but until that condition is documented, this is just additional data in the chart. Coders or clinical documentation improvement specialists may use these reports to formulate queries for clarification, but they do not support coding.

Examples of some common scenarios:

Patient presents with uncontrolled diabetes. The coder can only code E11.9, diabetes, not otherwise specified. There is no code for uncontrolled diabetes. There are separate and distinct codes for diabetes with hyperglycemia (E11.65) or diabetes with hypoglycemia (E11.64X, last digit depending on with or without coma), but the condition must be captured by the provider. The coder cannot look at the lab results and code anything more than unspecified until it is clarified by the provider. It matters: Both E11.65 and E11.64X are in a higher weighted risk adjustment category, which impacts the patient’s risk adjustment factor, which will impact an individual provider MIPS score, as well as potentially impact payment.

Patient has urosepsis. The coder can only code N39.0 for unspecified urinary tract infection. Is that what is meant, or does the patient have septicemia or sepsis? Without clarification, serious conditions may not be able to be coded, not only impacting the risk adjustment factor, but also failing to reflect the true acuity of the patient being treated.

GI bleed and anemia. The coder is able to code K92.2 for gastrointestinal hemorrhage and D64.9 for unspecified anemia. There is a more specific code, D50.0, for iron deficiency anemia secondary to blood loss; however, this cannot be chosen unless the provider captures the cause and effect relationship of the two conditions, i.e., anemia secondary to GI bleed.

Diagnosis coding is crucial first and foremost to accurately reflect the severity of the patients being treated and served. However, accurate and complete diagnosis coding also can significantly impact the revenue cycle and financial health of the organization. As value-based purchasing grows, so will the importance of coding accuracy.