When Does The PI Clock Start Ticking? The Answer, Part 2!

I analyzed the first of two PI clock scenarios in my last post. They are not always as obvious as they seem. Now let’s look at the second case:

A young male is involved in a motor vehicle crash and strikes his head. He enters your trauma center at exactly midnight as a trauma activation. Head CT shows a 7mm epidural hematoma with no shift and no effacement. GCS is 15, and the neurologic exam is completely normal. He is admitted to the SICU for neuro monitoring and is scheduled to have a repeat CT scan at 06:00. The scan shows significant expansion of the hematoma, with midline shift and ventricular effacement. He is taken to the OR for craniotomy by neurosurgery at 6:55.

This one is very similar to the first, except there is no indication to go to the OR at initial presentation. But about 7 hours later, he is in the operating room. So the PI trigger occurs, right? That’s more than 4 hours!

Not so fast! Let’s analyze this a bit more. Everything seems to be going well until the 6 AM CT scan. If the patient’s condition is unchanged, the earliest possible time the change in his head could have been recognized was shortly after 6:00. So the patient was actually in the OR less than an hour after the problem was recognized, right?

Not quite so fast again. The trauma PI program still has to examine the entire process from arrival until operation. Here are the questions that need to be answered:

  • Was neurosurgery involved in the initial evaluation in a timely manner?
  • Was the patient admitted to an appropriate inpatient unit?
  • Did appropriate monitoring occur?
  • Did any change in exam occur that could have suggested the hematoma was changing?
  • If so, did nursing and physician staff act appropriately with that information?

Bottom line: If everything went according to plan, and there was no change in exam or vital signs through the repeat CT scan, then this is an exemplary catch, and instead of sending the usual trauma PI nasty-gram to neurosurgery, they should receive a congratulatory note for providing such excellent service!

All too often, the trauma program just routinely sends out these “nasty-grams” without doing any further analysis of the data. And in cases like this one, the work involved in responding is just a waste of time. 

General rule: If the actual time noted for one of these time-sensitive filters is very, very long (e.g. delay to laparotomy of 62 hours), then look at it very closely. Did someone actually sit on a bleeding spleen for nearly three days, of was the patient doing well and suddenly failed nonoperative management? I think you know the answer.

And don’t forget to send out a few love letters to the other services for work well done from time to time! They probably cringe when they see trauma PI notes, since they always seem to imply something bad has happened.

When Does The PI Clock Start Ticking? The Answer, Part 1!

In my last post, I presented two potential performance improvement (PI) cases. I asked for your input as to when the clock should actually start for the 4-hour craniotomy/craniectomy rule. Today, I’ll give you my answer to the first case.

Lets look at it again:

A young male is involved in a motor vehicle crash and strikes his head. He enters your trauma center at exactly midnight as a trauma activation. Head CT shows a 12mm epidural hematoma with 8mm midline shift and ventricular effacement. GCS was 14 on arrival, but has declined to 12 by the time you leave the CT scanner. He is taken to the OR for craniotomy by neurosurgery at 4:15.

This one looks straightforward, right? But not so fast. The crani occurred more than 4 hours after arrival. Isn’t that a violation of the 4 hour filter? But did you know he needed an operation when he arrived in the ED? No! GCS and exam were reasonable, so the clock starts once the CT scan finishes, even if the surgeon doesn’t see them at that time. Why then? because the 4 hour rule is testing all of the following:

  • Whether a physician was present in CT and recognized what was on the images (not required, but reviewed if there was one there)
  • How long it takes for the radiologist to get the images
  • How long it takes for the report to be done
  • How quickly the surgeon or emergency physician review the report
  • How long it takes to contact the neurosurgeon
  • How long it takes them to see the patient and decide they need an operation
  • How easy it is to get this emergency case to the OR suite
  • How long it takes for anesthesia to do their assessment and get the patient into the room
  • How long it takes the OR team to be ready to cut

Lots of stuff! So if the scan finished any later than 12:15 am, this filter gets triggered. But hold on! In my opinion, 4 hours is a long time to wait for an emergent problem like this large epidural. Even if the scan finished at 12:30, the 4 hour rule is met, but why did it take so long to get the operation started? I’ve seen cases like this where the incision was started less than an hour after the patient arrived in the trauma bay!  Some of these cases need review even if they appear to meet the time limits.

Bottom line: Case #1 – the clock officially starts when the proof of clinical injury has been provided. This could be an abnormal physical exam, a CT scan, a critical lab test draw, a phone call from a concerned nurse, etc. The clock doesn’t necessarily start when the patient rolls in the door, unless you have some kind of weird superpowers!

I’ll review and analyze the second case tomorrow.

When Does The PI Clock Start Ticking?

This is a question that comes up frequently in trauma performance improvement (PI) programs. Several of the PI audit filters typically used at trauma centers include a time parameter. Some of these include:

  • Craniotomy > 4 hrs
  • Laparotomy > 4 hrs
  • OR for open fracture > 8 hrs (although this is now outdated)
  • OR for compartment syndrome > 2 hrs

The question that needs to be asked is: 2 or 4 or 8 hours after what?

Let’s consider the following scenario:

A young male is involved in a motor vehicle crash and strikes his head. He enters your trauma center at exactly midnight as a trauma activation. Head CT shows a 12mm epidural hematoma with 8mm midline shift and ventricular effacement. GCS was 14 on arrival, but has declined to 12 by the time you leave the CT scanner. He is taken to the OR for craniotomy by neurosurgery at 4:15.

And this one:

A young male is involved in a motor vehicle crash and strikes his head. He enters your trauma center at exactly midnight as a trauma activation. Head CT shows a 7mm epidural hematoma with no shift and no effacement. GCS is 15, and the neurologic exam is completely normal. He is admitted to the SICU for neuro monitoring and is scheduled to have a repeat CT scan at 06:00. The scan shows significant expansion of the hematoma, with midline shift and ventricular effacement. He is taken to the OR for craniotomy by neurosurgery at 6:55.

My questions for you:

  • When does the PI clock start ticking in each case?
  • What information do you need to review to make this decision?
  • Do you send a PI “love note” to the neurosurgeons in either case?

Share your thoughts on Twitter or by commenting below. I give you my answers in the next post.

Does Trauma Center Level Make A Difference In Treating Solid Organ Injury?

In the last two posts, I reviewed contrast anomalies in solid organs, specifically the spleen. Today, I’ll be more general and examine a recent paper that compared management and outcomes after the other major solid organ injury, liver, at Level I vs Level II trauma centers.

There are several papers that have detailed overall differences in outcomes, and specifically mortality, at Level I and II centers. Some of these show outcomes that are not quite as good at Level II centers when compared to Level I. On paper, it looks like these two levels should be very similar. Take away research and residents, and maybe a few of the more esoteric capabilities like reimplantation, and aren’t they about the same?

Well, not really. They can be, though. Level I criteria are fairly strict, and the variability between difference Level I centers is not very great. Level II criteria are a bit looser, and this allows more variability. Many Level II centers function very much like a Level I, but a few are only a bit higher functioning than a Level III with a few extra surgical specialists added in.

A paper currently in press used the Michigan Trauma Quality Improvement Program (MTQIP) data from all 29 ACS verified Level I and II centers in the state (wow!). Six years of information was collected, including the usual demographics, outcome data, and management. A total of 538 patients met inclusion criteria, and this was narrowed down to 454 so statistical comparisons of similar patients could be made for Level I vs Level II centers.

Here are the factoids:

  • Mortality was significantly higher in Level II centers compared to Level I (15% vs 9%) and patients were more likely to die in the first two days, suggesting hemorrhage as the cause
  • Patients were more likely to die in the ED at Level II centers, despite a significantly lower Injury Severity Score (ISS)
  • Pneumonia and ARDS were significantly more likely to develop in Level II center patients
  • Level II centers used angiography less often and took patients to the OR more frequently
  • Level II centers admitted fewer patients to the ICU, but ICU admission was associated with significantly decreased mortality
  • Complications were fewer at Level II centers, but they were less likely to rescue patients when they occurred

Bottom line: Level I and II centers are supposed to be roughly the same, at least on paper. But a number of studies have suggested that there are more disparities than we think. Although this paper is a retrospective review, the sheer number of significant differences and its focus on one particular injury makes it more compelling.

So what to do? Tighten up the ACS Orange Book criteria? That’s a slow and deliberate process that won’t help our patients now. The quickest and most effective solution is for all centers to adopt uniform practice guidelines so they all perform like the highly successful Level I programs in the study. There are plenty of them around. If you are not yet using one, I urge you to have a look at the example below. Tweak it to fit your center. And use your PI program to trend the outcomes!

Related post:

Reference: Variability in Management of Blunt Liver Trauma and Contribution of Level of ACS-COT Verification Status on Mortality. J Trauma, in press, Dec 1, 2017.

Natural History of the Splenic Blush

In my last post, I described the two types of solid organ “blushes.” I also described my thoughts on the natural history of these findings. Now, a multicenter study on the natural progression of the splenic “blush” has just been published. I found this paper very interesting, because it challenged some of my own existing beliefs. But once I read it, my enthusiasm faded.

The Western Trauma Association sponsored a multicenter (17 Level I and II centers) review of data collected prospectively over an unspecified period of time. Patients were excluded if their injury was older than 24 hours, if they had a previous splenic injury, and if they had any number of diseases or hereditary conditions that might affect the spleen. Strict definitions of nonbleeding and actively bleeding injuries were applied, and detailed information on intervention and outcomes was collected.

Here are the factoids:

  • 200 patients were enrolled from 17 centers, but the paper does not state how long that took
  • 20% were low grade (1 or 2) and 80 % high grade (3-5)
  • 29% had a pseudoaneurysm, and 83% showed extravasation, which means that several patients had both
  • 15% underwent early splenectomy, 59% underwent angiography, and 26% were observed
  • For those with initial angiography, 6% had repeat angio and 7% eventually underwent splenectomy
  • Of those were were initially observed, 9% had delayed angio and 8% underwent splenectomy
  • Based on a read by an expert radiologist, an actively bleeding injury was associated with a 41% splenectomy rate
  • The authors conclude that the majority of patients with spleen injury with pseudoaneurysm or extravasation are managed with angio and embolization and that splenectomy remains a rare event (??)

Bottom line: This paper just doesn’t do it for me. The biggest problem is that it is what I call a “we do it the way we do it” study. It examines how 17 different centers evaluate and treat patients with significant splenic injury. There was no guidance or guideline on how to treat, so they each did it their way. And the number of patients was small.

They don’t tell us anything about the use or effectiveness of angio by grade. Or whether the specific hospitals routinely rely on angio rather than just going to the OR for high grade injuries (typically if angio response times are long).

Unfortunately, this paper gives the appearance of containing a lot of interesting stuff. But a 15% initial splenectomy rate is not a “rare event” in my book. Everything published here is at odds with what I’ve observed over the years for centers with well developed management guidelines and easy access to angio (< 5% splenectomy rate in hemodynamically stable patients with nonoperative management).

My recommendation is to send all stable patients with pseudoaneursym and/or extravasation to angio immediately! Yes, some will have nothing found by the time they get to angio, and you’ll have to come up with a plan at that point. But most have something wrong, and it won’t stop until it’s been plugged up (or your patient bleeds to death, whichever comes first)!

This article has all the right buzzwords: multicenter, prospective data, etc. But it’s already been moved to my recycle bin. 

Related post:

Reference: Natural history of splenic vascular abnormalities after blunt injury: A Western Trauma Association multicenter trial. J Trauma 83(6):999-1005, 2017.