Difference between revisions of "Talk:Accident analysis"

Latest revision as of 11:47, 6 May 2009

Suggestions from Richard Mundy

Were I now to be analysing this accident at work, I would now think about this in the context of the 'swiss cheese model' (http://en.wikipedia.org/wiki/Swiss_Cheese_model). The idea behind this is that between each hazard (falling whilst descending a pitch) and its possible worst consequence (fatality) lie a number of barriers (slices of swiss cheese), each of which has weaknesses that prevent them all from preventing all consequences all of the time (holes in the cheese). We are often invalidating some of these barriers, but it is when we simultaneously circumvent all of them that the fatality is suffered.

Furthermore, we can subdivide the barriers between those preceding the main event (falling) from after it (those between the fall and death). Barriers can also be subdivided into 'hard' barriers (engineered ones) and 'soft' barriers (human/procedural ones). Here we prefer to rely on barriers preceding the main event (i.e. preventing the fall rather than escalation of its consequences). 'Hard' barriers tend to be more reliable than 'soft' ones.

The barriers I have identified (there are more that could be added) in some order:

HAZARD: descending pitch - Equipment designed safely (STOP was able to be deployed on something too weak to support bodyweight) - Adequate equipment deployed (caver was otherwise unattached to the rope) - Caver trained to a competent degree (caver failed to use his equipment properly) - Caver adequately safety minded (caver failed to check himself at the pitch head) - Party checking one another's equipment deployment (nobody else intervened) MAIN EVENT: fall - Equipment designed safely (helmet and oversuit prevented some injuries, but allowed others) - Party intervened adequately (companion raised alarm fast, acted to conserve body heat) - Emergency response adequate (fast, coordinated effort, competent rescue) - Subsequent care good (hospital acted well) CONSEQUENCE: fatality

The conclusion to draw from this would be that the most effective thing we could do would be only to use harnesses to which all points with loops big enough to attach STOPs are strong enough to support the entire bodyweight. Sadly I doubt that CUCC, the gear industry, or cavers in general feel strongly enough about safety to make this happen.

The next best thing would be to attach a descending caver in a second way, perhaps a shunt attached separately to the harness. There would be nothing to stop this being wrongly attached too, but the odds are reduced. This would make descending slower, but is common practice in the professional world (e.g. double rope technique).

If we reject each of these 'hard' interventions, we will therefore be relying on 'soft' barriers. Perhaps a caver needs to go through training whereby they correctly kit up on their own, pass rebelays etc. a certain number of times at heights below 5m before being let loose on bigger pitches. Perhaps all cavers should be drilled in checking a list of things before each pitch, or before the first pitch of the day. Perhaps all parties caving should have such a checklist they run through with one another, or just with novices.

Luckily the barriers after the event functionned well. His helmet perhaps prevented instant death, the immediate response perhaps prevented hypothermia, the emergency response perhaps prevented some internal bleeding, etc. Was it lucky that this was the case though? Could the mobile phone at base camp equally have been off? Could the top camp one have had no battery? What if there had been no CUCC member speaking German?

There are probably hundreds of things that could be done. Again, were this a work situation we would probably ban caving and design an ROV to do it!

Perhaps the mode of thinking above gives you some idea of what could/should be done? For me, ensuring novice competence and obliging others to check novices in set ways are probably the most palatable and effective things to persue,

Richard

Suggestions from Mark Dougherty

Some clear issues emerge conserning management of the hypothermia. This was a great success, but was achieved in the face of some considerable adversity - candles going out, matches running out, sleeping bag hard to put on etc. Equipment needs to be re-thought here. What EVERYBODY should be carrying (in my opinion) is:

• nightlight candles
• small folded sheet of tin foil to make a wind shelter for candles
• piezo-electric lighter
• hat/balaclava
• blizzard survival bag see: http://www.blizzardsurvival.com/product.php/100/blizzard-survival-bag The blizzard survival bags are apparently the business. Everyone should have one in their kit. Yes they cost, but I'm assured by people who ought to know that they are a real life saver.

The sleeping bag for casualty use needs to have a full length zip. Can someone handly with a sewing machine modify the bag we have (which being fibre pile is very good from the point of view of insulating even when damp)?

if a stove is availble (either brought from the surface or at a brew station in the cave), make sure a plastic water bottle comes with it. You can heat water to a little over body temperature and then put a hot water bottle inside the casualty's clothes next to their core. If I think back to Razordance 2007 this was a serious omission from the brew equipment at GLAD and could have made a substantial difference to survivability.

A roll of binbags will waterproof anything in an emergency....

There are obvious SRT training issues. Basically anyone who has not caved regularly with CUCC should be checked out for SRT competence by the leader or other experienced person before being allowed underground. No exceptions, then it's not embarassing to insist on it.

Suggestions from Mike the Animal

• Keep a gas fueled emergency stove at top camp, one with a piezo igniter. Make sure its kept in a waterproof container (along with matches and a couple of ciggy lighters). I got one quite cheap off ebay and use it underground on Upper Flood digging trips, works fairly well [1], its a bit fiddly to control, possibly temperature related, but holding the gas canister in my hands works.

• I've used an Ortlieb dry bag when canoe camping as a stuff sack. The ones I've used are fairly long and thin, so will go into a tackle bag. Actually, one of mine has a tiny hole but would be fine with a thick bin liner inside - air pressure should mostly keep water out anyway; I'm happy to donate it.

suggested "buddy checking" practice

Thanks for doing this, Kathryn. It looks accurate and succinct. The only good thing about an accident like this is what we can learn from it, so I'm glad to see it recorded well.

As for suggestions for change: I spoke with Robert Seebacher after the incident, and he was unequivocal about what he thought should change. He said that on every pitch head of every trip, he always has someone look and check all is well (mallion gate, stop krab gate and attachment, harness straps doubled back), and checks theirs at the same time. This was a partial cause of the accident; I watched as "Caver A" tested his stop, but I did not actually look at his central mallion at the pitch head. If I had, it would have been obvious that Caver A's stop had not been moved to his mallion.

Although this would be somewhat of an inconvenience, I strongly recommend that the club adopt this visual inspection ritual, or at least include it in training. I don't know if the number of unscrewed-mallion incidents has decreased since I left, but that's another reason I think checking is good practice.

I can add this to the "Analysis and recommendations" section, but wanted to put it here for discussion first. What does everyone think of this? -Aaron 03:24, 1 May 2009 (UTC)

Comments from Mark D via email:

What Aaron says about buddy checks (and his report of a conversation with Robert on this subject) is very interesting. We instituted a similar policy in the Swedish Federation some years ago along these lines, with a buddy check before the first pitch. However, I don't think it's realistic to have a check before every pitch because this would tend to slow progress.

Part of the issue is that taking SRT kit off and back on again is a safety risk, just as moving the descender back and forth from the central maillon to a gear loop is also a safety concern. Several aspects immediately spring to mind: (a) There is obviously the risk of loading the descender when it is on the gear loop. (b) Re-installing the descender might be done wrongly, for example crab not done up, put on backwards etc. (c) on the ascent, if you need to change over to going down in an emergency there is an additional risk of dropping the descender mid pitch, or finding that you forgot it in a bag which is not attached to you. (d) Similarly there have been cases of people abseiling down to the bottom of a cave and finding they left their top jammer behind at the surface. That hasn't necessarily happened on EXPO but it's definitely happened to cavers before.

My suggested policy for the future is therefore as follows:

a) SRT kit is either "on" in its entirety or (by definition) it's "off" and a new buddy check needs to be made before getting back on a rope. SRT kit to be "on" whenever on a rope.

b) Buddy check performed whenever SRT kit put "on".

c) Central maillon tightened sufficiently so that it can't work loose (a 7mm long maillon makes an excellent spanner for this purpose).

d) Taking SRT kit on and off all the time to be avoided where possible because of the extra risks involved.

Final food for thought. Buddy checks are much easier and faster to perform if everybody configures their gear in the same way. There is a limit as to how regimented we can get, but it's maybe worth looking at standardising the basic layout of equipment.

Mark

Audience

What audience is this page aimed at?

• Laypeople (unfamiliar with caving jargon)
• Past/present expedition members
• New expedition members
• Cavers unfamiliar with the CUCC expedition
• All of the above

--OliverMadge 11:11, 29 April 2009 (UTC)

I'd say mainly the the second and fourth - i.e. anyone with any useful ideas. Edvin

One of the main reasons for doing it is that it is required by the CUEX expedition people, but I don't think they are that interested in reading it, they just want to know that we have done it (so I don't think it is important that it makes sense to non-cavers). Kathryn

Ah, did CUEX specify the section titles? I just renamed them slightly to make it clearer (to me at least). If so you might have to rename them back. --OliverMadge 11:52, 29 April 2009 (UTC)

The section titles were suggested by Mark D, but I'm sure he won't mind them being changed Kathryn

CUEX were certainly very keen on some sort of analysis. But I think it's a good idea to try and learn some lessons anyway; although keeping it as un-officious as possible is always good! Edvin