By Robert S. Gold, MD
I have come to realize that if something can be misinterpreted, it will be misinterpreted. The application of the Agency for Healthcare Research and Quality's (AHRQ) Patient Safety Indicator (PSI) 15 is one of these situations.
Hospitals and surgeons don’t want to “look bad” and hospitals don’t want to get financially dinged for reporting these events, which contributes to misinterpretation. But part of it is because the verbiage in the guidelines is open to interpretation, so different organizations interpret the guidelines differently.
This PSI tracks events during surgical procedures that can hurt patients. It does not track whether the patient actually suffers harm from the event.
Medicare is overstepping its bounds by using this PSI to punish hospitals where events occur. Surgical oncology centers are more likely to have a surgeon accidentally lacerates or puncture an organ during very complex operative procedures than the average hospital, which sends its tough cases to these centers. It’s the nature of the beast. Yet Medicare doesn’t distinguish between the two facilities and the physicians who do these operations and save lives get identified as bad surgeons.
Version 5 guidance
Let’s take a look at the guidance provided by Version 5, which was just released. It starts by identifying those patients who should be included in the statistical analysis and those to be excluded:
Accidental punctures or lacerations (secondary diagnosis) during procedure per 1,000 discharges for patients ages 18 years and older. Excludes cases with accidental puncture or laceration as a principal diagnosis, cases with accidental puncture or laceration as a secondary diagnosis that is present on admission, spinal surgery cases, and obstetric cases.
Patients who come into the hospital because of a problem that occurred prior to the current admission do not count for the receiving hospital. This holds true even if a surgeon at the same hospital was responsible for the original puncture or laceration, but did not recognize it at the time of the procedure and thus did not report it.
But surgeons and the American College of Surgery take umbrage with this sweeping definition of patients to be included, and cancer centers are most under the gun for being reported as “bad actors.” So they look at the guidance from AHRQ and read:
If a puncture, tear, capsular laceration, enterotomy, colotomy, serosal laceration, “injury,” or other such event occurs due to patient-specific factors (e.g., the nature of the adhesions, the inflammation, the abscess, the tumor, or other conditions present during the operation) that you believe are routinely expected and inherent to the procedure performed, your documentation must clearly state that the event was inherent to the surgical procedure to avoid the incorrect reporting of a complication.
Appropriate and consistent use of the terms inherent, integral, or intended to describe nonaccidental lacerations/punctures and explicit documentation when intraoperative or postoperative complications occur will reduce requests for clarification.
You are also encouraged to document the reasons for describing an event as inherent (e.g., tumor disease encroaching on surrounding tissues, organs friable due to prior radiation treatment) to further clarify the documentation.
Here’s where the writers of the guidance make it easy for providers to misreport and mess up statistics.
The goal of the context of “inherent” or “integral” or “intended” is to identify:
- Surgical procedures that were part of the operation listed on the operative permit
- Additional surgery that might not have been mentioned on the operative was required to complete the case satisfactorily and that decision was purposely made during the case – not to repair something accidentally transgressed
Examples
Let’s look at a few examples. The surgeon planned to insert a feeding jejunostomy for nutrition in a patient with duodenal obstruction. The surgeon performed an enterotomy to insert the feeding tube. That is the operation and not an accidental event.
The planned operation, as described on the op permit, was removal of a benign adenomatous polyp that partially obstructed the sigmoid but could not be removed endoscopically. The opening of the sigmoid to get to that polyp was the operation and not an accidental event.
During a pelvic operation for uterine cancer the surgeon discovers that the tumor is directly invading the dome of the bladder. The surgeon determines that removing the dome will improve the patient’s changes for a good outcome. That entry into the bladder with resection was proactively intended and was integral to the ultimate goal of trying for a cure.
During a resection of an obstructive segment of transverse colon because of stenosis from chronic diverticulitis, the surgeon discovers that the tail of the pancreas is so intimately involved with the phlegmon that it would be safer to remove the tail of the pancreas en bloc with the left transverse colon. That part of the procedure was inherent in the goal of treating the pathology and curing the patient. It included intended resection of the tail of the pancreas.
On the other hand, during resection of a portion of an ovarian mass, the surgeon lacerated the hypogastric vein and the patient lost 350 cc of blood. The surgeon called in a vascular surgeon to repair the vein. This is a complication. Was it avoidable? Maybe not. Did it occur? Yes. Did it require intervention? You bet. Is it reportable? Yes.
During a hemigastrectomy for a gastrointestinal stromal tumor tumor, the surgeon placed a clamp on the tail of the pancreas and then needed to perform a distal pancreatectomy because of damage done to the pancreas. That was a complication. It was not planned and it was not part of the intentional extension of the original surgery at all.
Reportable events
We see some examples of reportable and nonreportable events also in the description of PSI 15. For example, if the case involved a few serosal tears during extensive lysis of adhesions from previous pelvic inflammatory disease and nothing happened to the patient, it’s not reportable as below:
Tears, punctures, or lacerations that are not inherent/routinely expected should be reported if they meet the ICD-9-CM definition of an additional diagnosis.
In cases of uncertainty, the indications of clinical significance may provide useful clues for clinical documentation specialists and coders and suggest the need to query the provider for clarification.
Some indications of clinical significance may include, but are not limited to, the following, as provided by University of Michigan Health System (Gwen Blackford, written communication, April 25, 2012):
- Prolonged the patient’s stay
- Required increased nursing care
- Required follow-up
- Required a blood transfusion
- Required surgical repair
- Required return to the operating room for repair
- Required a consulting provider
- Affected the patient’s course of treatment or recovery
- Resulted in damage
This is an eye-opener. This means that if a laceration to the serosa of the bowel during lysis of massive adhesions that was repaired did not meet Uniform Hospital Discharge Data Set (UHDDS) criteria (expanded for this purpose), that you don’t report the laceration or the repair – at all. And the event never happened – unless the patient is readmitted a week later with a perforation at the laceration site. Then you report it as the principal diagnosis.
Basically, we have rather good guidance. Minor events that are clinically insignificant don’t get coded. Significant events get coded – events that have significant clinical significance get coded, whether the surgery was tough or not. Some things happen that have no effect on anything – don’t report them.
Other situations are actually the operation itself or part of the extension of the operation in order to cure that patient. These are not accidents and you should report the additional surgical procedures but not 998.2 (accidental puncture or laceration during a procedure, not elsewhere classified).
Some things that happen aren’t planned and aren’t required to extend the surgery to cure the patient and have consequences, meeting UHDDS criteria. They are accidents and should be reported. Report code 998.2 and the additional operative codes.
We have to dump our concepts of “expected” for reporting of the ICD code of accidental laceration. If it happened and could kill the patient, it happened.
Oh, and look – many E codes that used to be identifiers are no longer included in that list – only 998.2.
With all that said, is it still fair to the oncologic surgery facilities? No. Can we do something about it? We should. The Powers That Be are thinking along those lines and that’s good.
Editor's note: Dr. Gold is CEO of DCBA, Inc., a consulting firm in Atlanta that provides physician-to-physician CDI programs including needs for ICD-10. Contact him at 770-216-9691 or rgold@DCBAInc.com.