Selection of PDX

Our coders will select the Principal diagnosis based on how "invasive" the testing is.  For example:  A patient comes in with vertigo and hematemesis.  For the vertigo- brain CT, IV meds, and  ENT consult were ordered.  Pt. diagnosed with a peri-lymphatic fistula.  For the hematemesis - GI consult, IV meds, serial H/H, and an EGD (mild gastritis found) were ordered.  MD stated there was no source found for the bleeding.   The coder selected the hematemesis based on the fact that an EGD was done. ("More invasive")   The CDS selected the per-lymphatic fistula because it was a higher DRG and felt both conditions were what occasioned the admission to the hospital.   In this kind of case the procedure does not drive the DRG.

 I am wondering if your coders use this thought process to select the Principal Diagnosis? 



  • edited July 2018
    There is no simple answer.   The clinical truth lies wrapped up in the UHDDS definition here.

    "The condition established after study to be chiefly responsible for occasioning the admission of the patient to the hospital for care."  is what you are looking for.

    There is no coding rule that I am aware of that require the coder to select the PDX which most corresponds to the invasive procedure.  I advise you to request the coder cite the official reference for that (and then post it here, I would like to see it).

    If I may however, it looks like the coder selected is still wrong.  If the EGD found gastritis and the patient was documented as having hematemesis....then gastritis with hemorrhage would be the pdx (as long as it was clear that gastritis was responsible for the presentation and not some other concurrent GI diagnosis).

    The Indexing in ICD 10 is pretty clear:

    Gastritis ....

    ....With bleeding: K29.71 (Gastritis unspecified with bleeding).  You could query for acute vs chronic if you wanted an even better code.

    And under the new guidelines, the MD really doesn't even need to say the bleeding was from the gastritis as long as their is no documentation in the record to indicate an alternate cause of the bleeding.

    The coder should not be selecting hematemesis as the pdx.

    Official Coding Guidelines:

    Codes for symptoms, signs and ill-defined conditions are not to be used as the principal diagnosis when a related definitive diagnosis has been established.  Although K92.0 Hematemesis is not a chapter 18 code it is functioning here as an ad hoc symptom that woudl be included in K29.71.  

    Another basic coding guideline is to code to the highest level of specificity available and K92 category is an "other category" while K29.71 is the specific diagnosis.  You should only use "other" categories when the documentation lacks the specificity to allow for a more specific diagnosis and in this case, you do have the specificity to follow the index to a more specific diagnosis. 

    By the way, the source of the bleeding does not have to be present on the EGD to report the code:

    Coding Clinic 3rd Quarter pg 27 (I think 2017)

     "A patient presents due to acute gastrointestinal bleed (GI). An esophagogastroduodenoscopy (EGD) was performed, which showed gastric ulcers. The physician does not link the bleeding to the ulcer nor is it documented that these conditions are unrelated. May we assume a relationship between the gastrointestinal bleed and the ulcer. How should we report gastric ulcer in a patient with gastrointestinal bleeding?
     It would be appropriate to assign code K25.4, Chronic or unspecified gastric ulcer with hemorrhage. As stated in the ICD-10-CM Official Guidelines for Coding and Reporting, (I.A.15)

     The classification presumes a causal relationship between the two conditions linked by these terms in the Alphabetic Index or Tabular List. Unless the provider documents a different cause of the bleeding or states that the conditions are unrelated, it is appropriate to assign the combination code for these conditions."

    Since that is true for Ulcer with bleed based on the "with guideline" I don't see why it also wouldn't be true for gastritis with bleed based on the "with" guideline.  (The indexing procedure in an ICD 10 manual work the same for either case).    

    Bear in mind there is a logic to this, as mild gastritis is NOT likely to yield an exact identification of the site of bleeding on EGD in many cases anyway.  A lack of a non-specific source is a finding routinely associated with bleeding chronic gastritis.  You could query for the providers confirmation (and perhaps you should) but if there is no other suspected source of the bleeding then the clinical truth is that it probably was the mild gastritis so this forms the basis of a sound query for linkage between two conditions).  

    Phew, that was a lot.  Still with all that being said, I am still not convinced that the GI diagnosis HAD to be pdx here.  It would come down more to the question of what was the providers biggest concern at the time the order was written.  Was it the vertigo or the hematemesis?  I believe either could work under the UHDDS definition at the time of the admission. 

    One last thing.  The coder may be using this:
     2017 Official Coding Guidelines
    “In the unusual instance when two or more diagnoses equally meet the criteria for principal diagnosis as determined by the circumstances of admission, diagnostic workup and/or therapy provided, and the Alphabetic Index, Tabular List, or another coding guidelines does not provide sequencing direction, any one of the diagnoses may be sequenced first.”

    Notice it says "diagnostic workup and or therapy" as part of the criteria for selection of the pdx here.  The coder may be stating that since the therapy is not equal, the defintion of "circumstances of admission" are not equal and therefore the "two or more conditions equally meeting" rule of coding is not at play and the GI diagnosis supercedes.

    I have seen this logic by my fellow coders many times.  Sometimes I agree with them, some times I do not.  Some times it is overruled by other guidelines.  For example, original treatment plan not carried out says, official coding guideline Section 2, F pg 104 says:

    "Sequence as the principal diagnosis the condition, which after study occasioned the admission to the hospital, even though treatment may not have been carried out due to unforeseen circumstances"

    Meaning that the circumstances of admission are more important than looking at what treatments were done.   You also have to dig into the timing and clock.  "Circumstances of admission" being what governs the condition that chiefly "occasioned the admission" means you should not be looking much past 24 hours into the stay for your "diagnostic therapy or treatment provided".  If it is much past 24 hours, then those diagnostic tests and treatments can hardly be counted as part of the circumstances which caused the admission. I fully disagree with my fellow coders who point to diagnostic tests or procedures done 3 days plus into the admission as proof that the two admitting conditions were not co-equal when going that deep into the stay is clearly not in accordance with the UHDDS time period of "circumstances of admission".

    I can't solve all this for you, but hopefully I have given you some things to think about.  

  • Thank you Allen, that was an excellent discussion!

  • great answer allen.

  • edited July 2018

       Thank you Allen for your detailed response.   Sorry for the delay- I had some IT issues getting back  in here.  Your response does give me a lot to think about!

      I agree - I think the coders are using this part of the PDX definition to support this decision making.

    In the unusual instance when two or more diagnoses equally meet the criteria for principal diagnosis as determined by the circumstances of admission, diagnostic workup and/or therapy provided. 

     The gastritis was unlinked from the hematemesis  in a query so that is why the hematemesis was used as the PDX.

      I am wondering if this "more invasive testing"  philosophy is based on denials or just an interpretation of this guideline?   And I was curious if this was unique to our Healthcare System or it was common among all coders. 

    Thank you for the great discussion on the circumstances of the admission. (the last paragraph)    I agree with you that we need to look closely at the first 24 hours to determine the circumstances of admission and what really occasioned the admission to the hospital.  It really is key in determining the thought process when the patient was admitted.   In this case I really thought that both conditions were of equal concern.  

     The next time I hear this argument I will ask for the official reference.  If I get a response I will definitely post it here!



  • Thanks for the follow up Melinda.

    I do find one thing unusual.  The patient has established gastritis, but idiopathic hematemesis with no other known GI diagnosis?

    Really?    I see a clinical problem here....
Sign In or Register to comment.