The Impact of Coding on Maternal Outcomes: Part II

Coding blood transfusions on maternity patients saves lives.

If you are a “frontline” coder buried in the minutiae of day-to-day coding, discussions about quality initiatives may not be in the forefront of your brain. We often hear of the importance of these programs in improving patient outcomes, and how critical “accurate and consistent” coding is. But how does that translate to the production coder trying to meet a quota? In part two of my series on severe maternal morbidity data, I will start to explore specific areas that impact the daily life of coders.

In last week’s article, I stated that the maternal mortality rate in the U.S. is twice as high as that of other high-income countries. Hemorrhage is one of the leading causes of perinatal death. Evidence has shown that most maternal deaths caused by hemorrhage were preventable, and occurred because of failure to focus and act on clinical signs of hemorrhage – which could have led to lifesaving interventions, such as blood transfusion.

As coders, it’s not really necessary – although it is interesting – to know the clinical signs the patient care team is examining to determine which treatment protocols are necessary. But I think there’s a certain level of knowledge we all need to have so that we understand why we are being asked to code certain procedures. In analyzing and trying to reduce the number of maternal deaths caused by hemorrhage, experts are looking at treatment measures that save lives. As noted, one of those treatment measures is blood transfusion.

The problem with blood transfusion data over the last five years is that many institutions have stopped collecting it. In a move directly related to the transition from ICD-9-CM to ICD-10-PCS, many organizations simply stopped coding blood transfusions. In the years leading up to the Oct. 1, 2015 implementation of ICD-10, many stakeholders gathered to identify documentation gaps in their medical records and then decide if the areas identified still required reporting. During that review, it was identified that blood transfusion coding did not impact reimbursement, and was not needed for any other purpose. It was also decided that the data could be obtained either from the blood bank or through hospital charges.

Unfortunately, data from the blood bank and hospital charges is not available to those who are analyzing data for the purpose of improving care. They have been relying on claims data, and I don’t think many people realized that before we made the move from the ninth to the tenth revision of the International Classification of Diseases (ICD). 

Why did we stop coding blood transfusions? The biggest reason I have heard is that it’s difficult to find all the data items needed to report the code. The code characters capture:

• Character 1 – Section (Administration);
• Character 2 – Body System (Circulatory);
• Character 3 – Root Operations (Transfusion);
• Character 4 – Body System/Region (at the time of implementation, this could have been a peripheral artery, central artery, peripheral vein, or central vein; since that time, the Centers for Medicare & Medicaid Services/CMS has removed the body part options for transfusion into a peripheral or central artery);
• Character 5 – Approach (until Oct. 1, 2021, the options were percutaneous or open, but the open approach was removed for FY 2022);
• Character 6 – Substance (type of blood product, such as packed red blood cells); and
• Character 7 – Qualifier (autologous or non-autologous).

Having used electronic medical records (EMRs) for several clients, I can attest that this documentation is not always tidily kept in one part of the EMR. The solution? Utilize clinical documentation improvement (CDI) programs to ensure that the type of blood product transfused is documented, and identify defaults for other code components based on common practice. This involves working with the obstetrical medical staff to determine the most likely scenario for transfusing blood in an OB patient. Most OB patients don’t have central venous access, but they have IVs. If a facility determines that the most common clinical practice in its institution is to transfuse blood bank (non-autologous) products via a peripheral vein, the hospital may develop an internal guideline stating that all blood transfusions will be coded using the body part for peripheral vein and the qualifier for non-autologous. 

Coding Clinic for ICD-10-CM/PCS published an article in the third quarter of 2014 that supports this advice. In that article, a question was presented about selection of the body part character for harvesting a saphenous vein graft. At that time, PCS had separate body part values for the greater and lesser saphenous veins (this has since been changed to a single body part value for the saphenous vein). In the published answer, Coding Clinic indicated that it was reasonable for facilities to develop internal coding guidelines for capturing the body part character of the specific saphenous vein harvested, based on common clinical practice. It should be noted that this guidance does not extend to the diagnostic side of this equation; Coding Clinic has published advice stating that it is inappropriate to assign a code for acute blood loss anemia based on clinical criteria, even if such criteria is written into a facility policy.

So, consider this my plea to coding professionals. If you aren’t assigning codes for blood transfusions on your maternity patients, please start.

Someone’s life could depend on it.

Programming Note: Join Kristi Pollard today on Talk Ten Tuesdays,10 Eastern, when she continues with part two of her exclusive four-part series on maternal morbidity and mortality.