All posts by TheTraumaPro

The Value Of In-House Call – Part 2

In my last post, I detailed an older study that did not show much of an impact from having the trauma surgeon in-house at all times. However, like many of the papers published on this over the years, it suffered from small numbers and questionable endpoints.

A group in the Netherlands sought to analyze everything they could find on the topic and perform a meta-analysis and systematic review. They scanned the literature beginning in 1976, the year that the ACS Committee on Trauma published the first resource criteria for trauma center verification. A total of 16 studies (RCTs and observational) that included information on over 64,000 patients were carefully selected for study. The endpoints of interest were in-hospital mortality and several process measures including lengths of stay and time to OR and CT.

Here are the factoids:

  • In-house mortality significantly decreased with in-house surgeons, with a relative risk reduction of 14% (from good quality papers, primarily published after 2000)
  • ICU length of stay was shorter with an in-house surgeon in four studies, longer in one
  • Hospital length of stay was shorter with the in-house surgeon in four studies, longer in two
  • Time to OR was significantly faster in seven studies with an in-house surgeon, but no difference was seen in five
  • Time to CT was shorter in one study and no different in four with the in-house surgeon

Bottom line: What does it all mean? We have been led to believe that doing a meta-analysis / systematic review can help us make sense of a group of papers with flaws such as low numbers, questionable design, or bias. This work shows that this is not necessarily the case.

Think of a  good meta-analysis as a set of eyeglasses focused on a selected body of literature. The blurry individual papers are grouped together and brought into better focus by the meta-analysis process. However, the final visual acuity is still determined by the overall quality of the individual research works.

If the overall quality is low, things will remain somewhat blurry even after meta-analysis. As individual paper quality improves, or the papers at least include some higher quality data mixed in with chaff, the overall clarity of the meta-analysis gets better and better.

In this meta-analysis, all papers included mortality information. There is enough there to show the association of an in-house trauma surgeon and lower mortality. But as with all association studies, it is impossible to say that the improved survival is due to the surgeon alone. There are many other factors that were not or could not be evaluated in the studies that might parallel the presence of the surgeon. And similarly with the process measures (LOS, time to resource use), we are generally seeing a preponderance of that show a positive effect. But it’s still not open and shut. 

I interpret this meta-analysis / systematic review as overall positive and supportive of having an in-house surgeon. It definitely dovetails with my own experience with in-house call over the past 38 years. I recognize the crudeness of the outcome measures selected, and our inability to quantify more subtle benefits. And we still haven’t fully figured it out the value, even after over 20 years of decent studies. This means we probably won’t ever fully know the answer since the system we work in continues to shift, potentially rendering the older information obsolete.

We will most likely continue with in-house call at highest-level trauma centers for the foreseeable future. In my opinion, and as is suggested by most of the literature, that is a good thing for our patients.

Reference: In-house versus on-call trauma surgeon coverage: A systematic review and meta-analysis, Journal of Trauma and Acute Care Surgery: August 2021 – Volume 91 – Issue 2 – p 435-444,

The Value Of In-House Call – Revisited Again

The value of in-house call for trauma surgeons has been contested for over a decade. Metrics for attending surgeon presence for trauma activations first appeared in the 2006 Optimal Resource document from the American College of Surgeons (Green Book). It called for the surgeon to ideally be present upon patient arrival, and no later than 15 minutes afterwards. This necessitated many trauma centers to mandate that the surgeons take in-house call so they could meet this standard.

As you might imagine, many were not happy about this. At Level I centers, the surgeons wanted to be able to rely on residents to help meet this requirement. The ACS was not too keen on letting them. So of course, people started doing research on the topic to prove their point of view.

I’m going to start off with an early paper on the topic from 2013. It was a rather sad initial attempt to show that surgeon presence didn’t make a difference. I’ll re-review that paper today, then move on to a more interesting one in my next post.

Of note: if you read just the abstract of this paper, you may come to the wrong conclusion! This is a perfect example of why you can’t just rely on the title or the abstract. Sometimes they cover up major flaws in the study.

This retrospective study primarily  of changes in patient mortality, as well as a few other length of stay (LOS) indicators  as the center changed over from having trauma surgeons who took call from home to taking in-house call. It involves only one trauma center in Lexington, Kentucky and covers two 21 month periods.

Here are the factoids:

  • There were roughly 5000 patients each in the at-home and in-house groups
  • Overall demographics looked identical, even though the authors thought they detected differences in age and ISS
  • Time in ED, ICU LOS, hospital LOS decreased significantly, and percent taken to OR increased in the in-house group. There was no change in mortality.
  • These patterns were the same in trauma activation patients, who were obviously more seriously injured.
  • The authors conclude that having an in-house surgeon does not impact survival, but can speed things up for patients throughout their hospital stay.

I have many problems with this study:

  • The statistical results are weird. Many of the allegedly significant differences appear to be identical (e.g. mean age 44+/-19 vs 45+/-19, hospital LOS 3 days vs 3 days). And even if the authors found a test that makes them look statistically significant, they are clinically insignificant. ICU LOS differences were measured in hours, and 25 hours was significant?
  • Attending presence “improved” from 51% to 88%. This means that they were not present in 1 of 5 trauma activations. This can easily overshadow any positive effect their presence may have had.
  • Mortality is too crude an indicator to judge the value of surgeon presence.
  • Lengths of stay can be due to so many other factors, it is not a valid measure either.
  • A retrospective, registry study has too few of the really critical data points

Bottom line: This paper is the poster child for why you MUST read the full paper, not just the abstract. If you had done the latter, you may believe that having an in-house surgeon is not necessary. Many papers prior to 2013 (of variable quality) have looked at this (poorly) and there is no consensus yet. But in-house call is a requirement for ACS verification if the surgeon can’t make it to the bedside of a seriously injured patients within 15 minutes. 

After observing trauma activations for 40 years, I know there is value in having an experienced surgeon present at the bedside during them. However, this value is very hard to quantify and every paper that has tried thus far has not looked at the right variables. And these variables cannot be assessed in a retrospective, registry type study. 

In my next post, I’ll look at a recent and better paper on the topic.

Reference: Influence of In-House Attending Presence on Trauma Outcomes and Hospital Efficiency. J Am College Surg 281(4):734-738, 2013.

The Tertiary Survey For Trauma: Residents vs APPs

This is the final installment of my series on the tertiary survey for trauma.  For years, this exam was performed by trauma surgeons or residents. However, over the years advanced practice providers (APPs) such as physician assistants and nurse practitioners have become more common in trauma. It is now commonplace for these providers to participate on the trauma service, perform procedures, and document examinations such as the tertiary survey.

But until now, no one has compared the accuracy of this exam when performed by a physician vs an APP. One would assume that the results should be the same, but as we’ve seen time and time again, common sense doesn’t always pan out. A group at the Royal Brisbane and Women’s Hospital in Queensland, Australia tried to answer this question using a retrospective review of their experience.

This busy trauma center admits about 2,250 patients per year, and began to employ clinical nurse consultants on the trauma service nearly ten years ago. Since there was no formal trauma curriculum for these nurses, they were required to complete the Trauma Nursing Core Curriculum (TNCC) or an equivalent prior to hire. The nurses were supervised by one of the trauma / emergency physicians.

For this study, 165 patients who underwent a tertiary survey by both an emergency medicine resident and a trauma nurse over a three year period were reviewed. The surveys were typically performed within 24 hours of admission to a ward bed or 24 hours before transfer from ICU to the ward. Typically, the resident and nurse tertiary surveys were performed within 30 minutes of each other to avoid any effects from injury progression.

All missed injuries were graded for severity by an attending physician using the Clavien-Dindo system. Here’s what it looks like:

And here are the factoids:

  • A total of 3,065 patients had a tertiary survey performed during the study period, but only 165 had it performed by both a resident and an APP
  • Based on their surveys, additional investigations were ordered in 35 patients, 14 by the trauma nurse, 11 by the resident, and 10 by both
  • Eight of 14 studies ordered by the nurse identified a missed injury, two of 11 studies ordered by the resident did, and two were identified in the studies ordered by both
  • Of the 12 identified missed injuries, the Clavien-Dindo (C-D) score was 0 in one, I in ten patients, and III (required surgery) in one
  • The nurses identified a higher number of missed injuries (10 of 24) than the residents (4 of 21) without significantly increasing the number of tests ordered

The authors concluded that performance of the nurses was similar to that of the house officers.

Bottom line: Maybe the authors were trying to be gentle on their residents. But it looks to me like the trauma nurses did a much better job of finding occult injuries. I wish the authors had broken down the C-D scores to see which group identified the score III patient.

To be fair, this study has some significant limitations. Out of more than 3,000 eligible patients, only 165 had a dual tertiary survey. So the sample may not be representative. But the results were impressive enough that I would speculate the results of a larger group may be similar.

So I think it is safe to assume that APPs (specifically nurse practitioners, but this can probably be generalized to physician assistants as well) can do a tertiary survey just as well as a resident. And possibly better!

Reference: Trauma tertiary survey: trauma service medical officers and trauma nurses detect similar rates of missed injuries. J Trauma Nursing 28(3):166-172, 2021.

The Tertiary Survey For Trauma: Does It Work?

Here’s the second part in my series on the tertiary survey for trauma. In my last post I discussed the basics, and in the next and final one I’ll review who can do it.

Delayed diagnoses / missed injuries are with us to stay. The typical trauma activation is a fast-paced process, with lots of things going on at once. Trauma professionals are very good about doing a thorough exam and selecting pertinent diagnostic tests to seek out the obvious and not so obvious injuries.

But we will always miss a few. The incidence varies from 1% to about 40%, depending on who your read. Most of the time, they are subtle and have little clinical impact. But some are not so subtle, and some of the rare ones can be life-threatening.

The trauma tertiary survey has been around for at least 30 years, and is executed a little differently everywhere you go. But the concept is the same. Do another exam and check all the diagnostic tests after 24 to 48 hours to make sure you are not missing the obvious.

Does it actually work? There have been a few studies over the years that have tried to find the answer. A paper was published that used meta-analysis to figure this out. The authors defined two types of missed injury:

  • Type I – an injury that was missed during the initial evaluation but was detected by the tertiary survey.
  • Type II – an injury missed by both the initial exam and the tertiary survey

Here are the factoids:

  • Only 10 observational studies were identified, and only 3 were suitable for meta-analysis
  • The average Type I missed injury rate was 4.3%. The number tended to be lower in large studies and higher in small studies.
  • Only 1 study looked at the Type II missed injury rate – 1.5%
  • Three studies looked at the change in missed injury rates before and after implementation of a tertiary survey process. Type I increased from 3% to 7%, and Type II decreased from 2.4% to 1.5%, both highly significant.
  • 10% to 30% of missed injuries were significant enough to require operative management

Bottom line: In the complex dance of a trauma activation, injuries will be missed. The good news is that the tertiary survey does work at picking up many, but not all, of the “occult” injuries. And with proper attention to your patient, nearly all will be found by the time of discharge. Develop your process, adopt a form, and crush missed injuries!

Reference: The effect of tertiary surveys on missed injuries in trauma: a systematic review. Scand J Trauma Resusc Emerg Med 20:77, 2012.

The Tertiary Survey for Trauma: The Basics

After a recent request, I’m re-posting a three part series on the trauma tertiary survey. Today, I’ll cover the basics. In the next two posts I’ll dig into how well it works and who can do it.

Major trauma victims are evaluated by a team to rapidly identify life and limb threatening injuries. This is accomplished during the primary and secondary surveys done in the ED. The ATLS course states that it is more important for the team to identify that the patient has a problem (e.g. significant abdominal pain) than the exact diagnosis (spleen laceration). However, once the patient is ready for admission to the trauma center, it is desirable to know all the diagnoses.

This is harder than it sounds. Physical examination tends to direct diagnostic testing, and some patients may not be feeling pain, or be awake enough to complain of it. Injuries that are painful enough may distract the patient’s attention away from other significant injuries. Overall, somewhere between 7-13% of patients have injuries that are missed during the initial evaluation.

A well-designed tertiary survey helps identify these occult injuries before they are truly “missed.” This survey consists of a structured and comprehensive re-examination that takes place within 48-72 hours, and includes a review of every diagnostic study performed. Ideally, it should be carried out by two people: one familiar with the patient, and the other not. It is desirable that the examiners have some experience with trauma (sorry, medical students).

Why 48-72 hours? Why not just do it when the patient leaves the ED, or when they arrive on the floor? Many occult injuries take time to show themselves. Swelling or bruising takes many hours to become obvious. And the patient may have distracting injuries and just won’t notice a sore finger or wrist that early.

And you can’t wait too long either! Otherwise the issue becomes a clearly delayed injury. A best practice is to require the tertiary survey be done within a specific window (24-48 hours, 48-72 hours, whatever works for your trauma team. Any injuries found in that time interval are not delayed diagnoses, since this process is designed to identify those pesky injuries. Any found after the time interval expires must go through a formal PI review at the primary and/or secondary levels.

The patients at highest risk for a missed injury are those with severe injuries (ISS>15) and/or impaired mental status (GCS<15). These patients are more likely to be unable to participate in their exam, so a few injuries may still go undetected despite a good exam.

I recommend that any patient who triggers a trauma team activation should receive a tertiary survey. Those who have an ISS>15 should also undergo the survey. Good documentation is essential, so an easy to use form should be used. Click here to get a copy of our original paper form. We have changed over to an electronic record, and have created a dot phrase template, which you can download here.

In my next post: Does the tertiary survey actually work?