Category Archives: SAR

Does that airbag really help?

Sorry for the long hiatus over the holidays. I wish I could say I was doing something fun and exotic, but instead I simply became the curriculum director at my current residency, which changed my workload quite a bit. That being said, back to the wilderness topics.

Since it’s winter, we are going to keep talking about avalanches. Since indications for resuscitation have been discussed before, now we will turn towards prevention. Two main ideas, first, avoid triggering an avalanche, and second, if you find yourself caught in one, try not to get buried. It makes sense, as the data from older studies is pretty clear that people who are buried die at a markedly higher rate than those who are not (52% vs 4%). What’s more, for those that are buried, the quicker they are found decreases their mortality, thus people buried less deep would likely have a higher survival rate.

Source: Hansi Heckmair/ABS

Enter airbags, which were invented to prevent this “critical burial” that causes increased mortality (critical meaning impairment of airways). They do this by basically making you much larger and more buoyant by inflating a large balloon that is strapped to your back. At least, that’s the theory. There wasn’t a lot of strong research devoted to them before implementation, and as they weren’t created by Roche, postmarketing research was lacking as well. Tie observer bias into this (people weren’t reporting near misses that didn’t involve airbags) and you are left with almost nothing of value to base recommendations on.

These authors wanted to fix that, so they did this study to determine the effectiveness of airbags based on preventing critical burial and mortality, as well as documenting frequency and causes of deployment failures. To do this, they looked at prior avalanche accident records from multiple countries, culling only worthwhile data that would show a difference between airbag users and nonusers. Thus, single victim events, small avalanches, or victims who weren’t seriously involved were removed in an attempt to reduce bias.

And once they crunched the numbers, they found out that airbags really do help. If you combine airbag failures with airbag inflations, the absolute risk reduction for critical burial is 29% (56%-27%), and the absolute mortality reduction is 17% (34%-17%). If you combine airbag failures with those not wearing airbags to begin with (why?), then risk reductions for critical burial and mortality are 35% and 23%, respectively. When you combine these values to adjusted mortality, you get a risk reduction of 11%, or a NNT of 9 for mortality with airbag use. Not too shabby.

Deployment failures occurred an alarming 20% of the time. Of these failures, 72% were attributed to operator error (not deploying them appropriately or incorrect maintenance). Slightly concerning, 12% of the failures involved destruction of the airbag during the avalanche. Of course, the absolute failure rate due to destruction or device failure is right at 5%.

So yes, if you’re going anywhere that there’s a risk of avalanches, you should wear an airbag. Also you should carry a beacon. And, like most other life saving measures, be they medical or technical, you’re only as good as what you do. Thus, read the instructions and know how to use it before you go out-of-bounds. However, this study did have a higher mortality of airbag users from prior studies (11% vs 3%), so don’t expect an airbag to make you immortal. Certainly, don’t do stupid things simply because you think you have a safety net (although it’s been shown that this doesn’t really occur). Of note, the usual problems with poor data due to non-standardized reporting as well as a low total number of victims apply to interpretation of this data.

The effectiveness of avalanche airbags
http://www.ncbi.nlm.nih.gov/pubmed/24909367

What leads to rock climbing rescues?

Rock climbing is a popular sport, but most rock climbing areas are either remote or sparsely populated with climbers, leading to poor data collection with regards to rescue events.The US National Park Service has data that shows only 3% of SAR are for technical roped climbing, versus 48% for hiking.  Thus, preventive education for climbers, as well as preparation for SAR, used to need a lot of educated guessing about what types of incidents occur.

Thankfully, there is one place where the density of climbers and the number of SAR events are high enough to get decent data. That place is Boulder County, CO, home of the Rocky Mountain Rescue Group. The incident reports by the RMRG allow consistent data for rock climbing rescues that break them down into specific injuries and climbing practices that led to them. It’s worth noting that climbing SAR accounts for 19.5% of the total SAR for RMRG, which is quite different from what the NPS reports. This is likely due to ease of access to popular climbing sites next to a highly populated area.

It’s not surprising that more incidents happen in summer and autumn. Weekends accounted for slightly more than half of the incidents as well. The most common victim is a male aged 20-29. In order of frequency, the activities that led to injuries were technical roped climbing (including belay incidents), unroped climbing, bouldering, mountaineering, and lastly simply being a bystander when a rock fall occurs.

For roped climbing, most injuries and incidents were lead falls, accounting for more than a third of incidents, followed by belay accidents at slightly over 20%. Being lost or stranded comes next, and have nearly twice as many victims per incident as any of the other activations. Second fall, anchor failure, rock fall, and medical causes are small players in this data set, accounting for single digit percentages each.

Incidents involving belay or rappelling most commonly resulted from the ropes being used not being long enough to reach the ground. A smaller amount resulted from the belayer losing control of the rope. Many of the rest of the incidents are from ropes getting stuck, with only 1 coming from a knot coming untied.

Unroped climbing includes free climbing as well as scrambling. The distinction is that free climbers are often experienced, and scramblers are not. The authors point out that their data does not distinguish between the two, so level of experience is not factored in to their incidents. They don’t show the data, but report that unroped climbing most often leads to victims being stranded but uninjured. However, they then report that unroped climbers are the most common climbers involved in fatalities, with 39% of fatal incidents.

Rock falls account for few injuries compared to the other activities, but their data demonstrates an important point. While rock falls are seemingly random, they most frequently occur during times of freeze-thaw cycling. Thus, in the spring they occur at lower altitudes, and this elevation goes up during the summer until the weather turns colder again.

Now that the activities have been identified, let’s turn to the injuries themselves. More than half (56.5) of the victims met by the RMRG had 1 or more injuries. Of those injuries, most were lower extremity injuries at 29.5%. This was followed by head injuries at 17%, and spinal injuries at 12.5%. Upper extremity, chest, abdomen, and “dislocated shoulder” (listed separately from upper extremity by the authors) were all less than 3% of the injuries each. Sadly, just under 10% (23/247) were fatalities. Of note, these are the suspected injuries listed by the RMRG, not actual identified injuries at definitive care, so be cautious with interpretation of severity of injury.

So what does this mean? Well, for the SAR guys, it means being prepared for more than just rescuing the lost, as many of them will be injured. However, preventing these incidents would have more health benefits. I’m not sure how to convince people not to climb without safety ropes, but the data shows that it is inherently riskier than all other forms of climbing. Also, making sure that your rope can get the climber all the way to the ground is quite important. When belaying or spectating, don’t stand in the fall line. Lead climbers should be extra careful based on this data set, as they are most at risk of injury.

Limitations are many. This is narrative data that is often presumptive as to injury, as well as missing many features such as length of fall, experience level, and other events leading to injury. There is also reporting bias, as it is likely that many minor injuries do not get reported to SAR. However, it is a robust set of data that hopefully can lead to changes in climbing education at popular areas.

Rock Climbing Rescues: Causes, Injuries, and Trends in Boulder County, Colorado

http://www.ncbi.nlm.nih.gov/pubmed/22727678

Props to the authors for giving a rather comprehensive list of definitions of the climbing terms used in their article. While many readers probably knew them, assuming that all readers do leads to decreased understanding of the article. I wish more authors would be as thorough with their explanations of possibly unfamiliar terms.

Predicting survival after avalanches

More than 150 people die each year after being buried in an avalanche, and mortality is greater than 50% for this condition. Unfortunately, a large amount of resources are used on patients who ultimately expire, so determining which ones are likely to survive can safe costs and allow better utilization of resources such as extracorporeal life support (ECLS) warming and air evacuation.

Of the 3 common causes of cardiac arrest after avalanche, only hypothermia is likely to have good outcomes. Trauma and hypoxia have poor outcomes. Most algorithms have providers stop resuscitation for severe trauma, and airways packed with snow. However, ascertaining hypoxia vs hypothermia is less obvious. Prior attempts used potassium >10 mmol/L or >12 as a surrogate marker for cellular death from hypoxia, but no other markers are used.

So these authors took 20 years of data from the North French Alps, which ended up being only 48 patients with cardiac arrest.  18 of them had ROSC pre-hospital, and only of those 2 were eligible for ECLS. 19 of the 30 without ROSC were also eligible for ECLS. In total, only 8 survived, 5 from the pre-hospital ROSC group, and 3 from the non-ROSC. Of the 8 survivors, only 3 had favorable neurologic outcomes.

All of these were patients with rescue collapse, that is loss of vital signs after extrication or transfer. 3 other patients with rescue collapse died however. Other indicators for survival in their analysis are the presence of a rescue pocket, K <4.3 (nobody survived above 4.2, but some nonsurvivors had levels below this), and coagulation disturbances. Interestingly, their data showed no overlap of prothrombin time between survivors and non-survivors, but they sadly did not give the values, only as ratios. Other values such as PaO2, PaCO2, lactate, and bicarb are not predictive.

Unfortunately, for such a long time period of collections, there were very few survivors. The retrospective nature also limits analysis. It does look like we need to reduce the cutoff for resuscitation from values of K from 10-12 mmol/L to a lower number (7?). Also, identification of coagulation abnormalities may help. Perhaps POC thromboelastograms may be a way to identify those that do not merit resuscitation.

Survival after avalanche-induced cardiac arrest
http://www.ncbi.nlm.nih.gov/pubmed/24971508

Search and Rescue: the Science

Source: René Kieselmann rmk (Bergwacht Schwarzwald e.V., BWS)When someone is lost in the woods, it often becomes the responsibility of the search and rescue teams to find them. It’s not a simple as one may think, however. The grid lines they follow need to be designed to increase the ability to find the lost person by maximizing the area able to searched while still having a better than average chance of detecting said lost person.

The grid lines are also known as Sweep Width (W), and the distance between them is simple if you have a chart. While the US Navy and US Coast Guard have well-defined values for maritime and aeronautical environments based on multiple studies, on land there are many more environment types and not enough studies have been done to create a value for each one.

That’s where this paper comes in. They attempt to describe a quick and dirty method of determining for the innumerable situations where a search and rescue operation has to take place in an area without prior calculations. The method by which they do this is fairly ingenious.

They did a total of 10 experiments in different environments at different times of the year. Each time, they measured the distance to first detection (Rd), the distance at which a located object could no longer be located (Re). They also measured a value known as the Average Maximum Detection Range (AMDR) which is an average of Rd and Re from the 8 compass points. Measurements were performed using high, medium, and low visibility objects.

What they found was interesting indeed. First, the differences between AMDR and Rd were so small that they were effectively equal. This is useful, as Rd is much easier to measure as it does not require equipment. Second, Rd had a good correlation with the W. This was true for all types of objects, but if they broke it down by high, medium, and low visibility they still had positive correlation. However, correlation wasn’t as strong for medium and low, but this might be due to sample size.

In the end, what they found was simply taking the Rd for any given object in any given environment, you could multiply by the correction factor to find W. The values they obtained were 1.773 for high visibility, 1.556 for medium, and 1.135 for low. In practice this would mean for a medium visibility object detectable at, say 20 m, then the detection index would be 31.12 m, which would how far apart the grid lines would be. The authors feel so strongly about their calculations that they recommend SAR teams to start using them, and without any evidence to the contrary, I’m inclined to agree. Sure, I’d like more studies before wholeheartedly endorsing it, but when the alternatives are basically guessing, this is probably better.

Use of the Visual Range of Detection to Estimate Effective Sweep Width for Land Search and Rescue Based On 10 Detection Experiments in North America.
http://www.ncbi.nlm.nih.gov/pubmed/24462331