PSI-9: Can a Postprocedural Hemorrhage or Hematoma be Unavoidable?

Original story posted on: January 25, 2021

The CDI approach to PSIs should be to scrutinize the indicators for the inclusion and exclusion criteria.

I believe that patient safety indicators (PSIs) serve a legitimate purpose. They are intended to “prevent potentially avoidable safety events that represent opportunities for improvement in the delivery of care.” This is a reasonable, even desirable, goal.

It is also legitimate to take action to avoid being penalized for PSIs when they are not valid. However, I object to some institutions and systems going through contortions to try to ward off all PSIs at the expense of the truth. If your hospital has a high postoperative wound dehiscence rate (PSI-14), maybe you should get a different type of antibacterial soap in the operating suite. If there are too many perioperative thromboembolic events, perhaps your policy for prophylaxis against venous thromboembolism (VTE) needs to be revisited.

I co-chair the clinical documentation integrity (CDI) education committee of the American College of Physician Advisors, and we publish educational materials for our fellow PAs. One of the committee members recently brought up the topic of PSI-9, perioperative hemorrhage or hematoma rate, to my attention, as his institution was developing materials for its own use. There is precious little written about this in the CDI literature, so I did some research and conferred with my coding and CDI posse. Here’s my take.

In order to trigger PSI-9, there must be a hemorrhage or hematoma due to a surgical procedure performed during an admission that requires a second PCS procedure to address the issue, and there cannot be any extenuating conditions from a list of coagulation disorders. The index procedure must be performed in the operating room, but the subsequent procedure can be at the bedside or in the OR. Many root operations are included in the list of procedures, including repair, drainage, occlusion, extirpation, and control.

I am not sure why it is titled “perioperative hemorrhage or hematoma,” because the included diagnoses are all “postprocedural” hemorrhages or hematomas. There is a whole set of intraoperative hemorrhage and hematoma ICD-10-CM codes, but they are not included in PSI-9. In fact, I think this introduces one of the elements of confusion. It is possible that some are conflating PSI-9 with PSI-15, abdominopelvic accidental puncture or laceration rate. The conditions that fall under PSI-15’s purview are intraoperative misadventures of puncture or laceration.

This makes a difference, because the criteria for whether or not to use a complication code provide some stipulations. The provider must document that the condition is a complication, there must be medical assessment or treatment, and there has to be a cause-and-effect relationship between the condition and the procedure performed and/or care administered. The criterion that I see institutions attempting to utilize as a loophole is, “was the condition an unexpected outcome or occurrence resulting from the surgery and/or from a preexisting condition?”

Related to PSI-15, people often talk about documentation of “unavoidable,” “integral,” or “inherent to” as verbiage that constitutes exclusion from the indicator. If there are dense adhesions and an enterotomy is inadvertently caused, was it unavoidable? If there is hemorrhage because the neoplasm is excessively vascular, the bleeding is inherent to the pathology, and not directly caused by the surgery. If a patient undergoes a thoracotomy, an iatrogenic pneumothorax is expected, and is excluded from PSI-6.

Surgery 101 mandates assessment for hemostasis at the end of a procedure. If hemostasis, or control of bleeding, is not accomplished, you find the culprit and ligate or cauterize or staple or suture it until there is control. There are rare circumstances in which you cannot achieve control, and the patient is packed and either left open or closed with the intention of reoperating, like if the liver or spleen is extremely damaged after a trauma. You would not code this with a postprocedural hemorrhage or hematoma; it is a function of the original trauma, not a complication of the surgery, per se.

Similarly, if a patient is coagulopathic, even with apparent hemostasis in the operating room, there is the potential for oozing or bleeding post-procedurally. This patient would not be captured in PSI-9, because although there is a bona fide postprocedural hemorrhage/hematoma, the fact that there is a bleeding disorder due to circulating extrinsic anticoagulants (D68.32) – or perhaps D68.4, Acquired coagulation factory deficiency (e.g., due to liver disease) – would be an exclusion from the quality metric.

The example offered for our educational materials was a small bowel resection and anastomosis that was documented as being technically challenging and resulted in a hemoperitoneum and drop in hematocrit, two days post-operatively. Since the original surgery was so difficult and this might have happened to even the most experienced and technically proficient surgeon, shouldn’t the postoperative hemorrhage be considered integral, and not be picked up?

My answer is “no,” this patient still falls into PSI-9. No surgeon is perfect, and there are patients who have friable tissues or who don’t follow the surgeon’s instructions to avoid heavy lifting. A low-level incidence of postoperative hemorrhage or hematoma is to be expected. If any provider would have the same complication under the same circumstances, then you could expect the incidence to be reflected across everyone’s metrics.

In fact, there are benchmark data tables found in the Agency for Healthcare Research and Quality (AHRQ) Resources (AHRQ PSI Benchmark Data Tables). The overall rate of perioperative hemorrhage or hematoma is 2.25 per 1,000 admissions. If a hospital had a rate of 0, one might suspect tomfoolery.

One of my concerns is that if a second procedure must be undertaken to address a hemorrhage or hematoma, a diagnosis that supports the medical necessity of the procedure must be made. Bleeding from the ileocolic artery after a small bowel resection, requiring repair, can’t be coded as a “laceration of the artery” from the Trauma section of ICD-10, because it is not traumatic. You can’t use the reason for the original small bowel resection as the justification, because there no longer is a small bowel obstruction at the time of the reoperation, and the intent of the operation was repair of the artery. To tell the story, you need a diagnosis. The only logical diagnosis is I97.620, Postprocedural hemorrhage of a circulatory system organ or structure following other procedure. If they couldn’t identify a specific affected artery, then you would utilize K91.840, Postprocedural hemorrhage of a digestive system organ or structure following a digestive system procedure, instead.

Quality metrics are always a balance between actual quality issues and documentation errors, which give the perception of a quality problem that doesn’t exist. Clinicians should identify areas of opportunity for improvement in the delivery of care – and improve them. The most important thing to do to prevent triggering PSIs is to eliminate true quality issues, which are under the control of the hospital or surgeon.

So, the CDI approach to PSIs should be to scrutinize the indicators for the inclusion and exclusion criteria. Make sure your providers are documenting in alignment with the indicators, such as not using the word “postoperative” just to signify a temporal period. Familiarize them with exclusion criteria, and ensure that they are documenting them in a codable format, like D68.32. Make sure your queries to ascertain whether there was a complication or cause-and-effect are clear and non-leading.

But please, don’t try to get every postprocedural hemorrhage or hematoma wiped from the record to avoid a PSI-9. It isn’t ethical, it isn’t safe, and it defies clinical documentation integrity. Our goal should be to have providers practice excellent medicine and ensure that their documentation reflects it, and the quality metrics and reimbursement will fall where they belong. Let this PSI serve its purpose.

Programming Note: Listen every Tuesday when Dr. Erica Remer co-hosts Talk Ten Tuesdays 10 a.m. Eastern with Chuck Buck.

Erica E. Remer, MD, CCDS

Erica Remer, MD, CCDS has a unique perspective as a practicing emergency physician for 25 years, with extensive coding, CDI, and ICD-10 expertise. She was a physician advisor of a large multi-hospital system for four years before transitioning to independent consulting in July 2016. Her passion is educating CDI specialists, coders, and healthcare providers with engaging, case-based presentations on documentation, CDI, and denials management topics. She has written numerous articles and serves as the co-host of Talk Ten Tuesdays, a weekly national podcast. Dr. Remer is a member of the ICD10monitor editorial board, a former member of the ACDIS Advisory Board, and the board of directors of the American College of Physician Advisors.

Related Stories

  • A Question a Day will Keep the Queries Away: Acute Blood Loss Anemia
    I conceived a new project that triggered last week’s TalkBack segment on the Talk Ten Tuesdays weekly Internet radio broadcast, and I would love your feedback on the concept. My vision is to design a comprehensive clinical documentation improvement (CDI)…
  • Knowledge is Best When it is Dosed Daily
    I'm making a pitch for reciprocal education. EDITOR’S NOTE: Dr. Erica Remer reported this story live during a recent edition of Talk Ten Tuesday. The following is an edited transcript of her reporting. To set the stage, I am currently engaged…
  • Avoid Treating Patients Like Check Boxes on a List
    Quality of care cannot be achieved without good data. 2020 has brought us many looming changes in healthcare. Unfortunately, we have already seen increased reporting criteria, more administrative burden, and many payment cuts. As we address all of these, not…