Bubble model vs. Gradient Factors redux

[Dr Simon Mitchell]

Hello,

There has been a lot of discussion in this thread about an important principle exemplified by this question:

Is it not possible to test deep stops as a 10/85 against a 50/85, and the TTS is simply what it needs to be?

Yes, of course it is possible, but it is not particularly useful - not if you are seeking the 'truth in the universe' about optimal decompression strategy.

Lets say you believe in deep stops, and that's why you are interested in 10/85. And lets say that the 10/85 profile for a hypothetical dive prescribes 85 minutes decompression compared to 70 minutes for the 50/85 profile. What would a comparative study of these two profiles tell you?

Could it tell you that one profile is associated with better outcomes than the other? - YES it could.

And lets say your 10/85 profile has better outcomes, does that prove that deep stops are a good idea? - NO it doesn't

It doesn't because the 10/85 profile prescribed a longer decompression and you can't separate that from the effect of the stop depth distribution.

OK, you may say you don't care - all you were interested in was whether your 10/85 profile was better than 50/85, but think about that for a minute. Its not a smart position for an analytical decompression diver to be taking. The question you should be asking yourself is "OK, my favored 10/85 profile says I have to spend 85 minutes decompressing; so am I sure that the distribution of stops prescribed by 10/85 gives me the least risk for THAT length of decompression?"

Let me put it another way. By choosing 10/85 you have effectively chosen to do 85 minutes of decompression. Surely the only question that should now occupy your mind is: "if I'm going to do 85 minutes of decompression, what is the best way to distribute the stops?"

So, lets say you find that 50/70 gives you exactly the same length of decompression as your 10/85 profile, but with slightly less emphasis on deep stops and slightly more emphasis on shallow stops. Surely THAT (or something like it) is the comparison you should be making. Then you can decide whether your stop depth distribution over the duration of decompression you had decided to do is better (or not) than a different distribution of stop depth.

This is the crucial concept of decompression efficiency - the least risk for the same amount of decompression time. The best way to see it is not "what is best VPM vs ZHL or GF this vs GF that", but rather "I've chosen to do X minutes of decompression; what distribution of stops within those X minutes gives me the safest ascent?"

Getting back to the wider subject, the early promise of deep stops and bubble models was that they were more efficient. They would get you out of the water in the same or even less time, with less bubbles and greater safety. The concept was embraced with great enthusiasm in the early 2000s based entirely on theoretical attraction and no data. Subsequently, every human study comparing an approach emphasizing deep stops (eg bubble models or ratio deco) has shown the opposite. There is a strong signal in the literature that if you over-emphasize deep stops any advantage of reducing supersaturation in fast tissues early in an ascent is outweighed by the disadvantage of greater supersaturation in slower tissues later after the dive - a construct that I suspect is supported by those results in the Powell study (graphs presented earlier in this thread).

It seems likely that bubble models and decompressions that emphasize deep stops are not the most efficient approach to decompressing. Doesn't mean you should not use them of course, and the material differences in risk compared to other approaches with the same duration may be small. Moreover, I cannot tell you in an evidence based way how far to back away from the bubble model style deep stops that we all embraced in the early 2000s. It is almost certainly not as far as raw Buhlmann, but that information is still to come.

Simon M

 

[Dr Simon Mitchell]

Hi EFX,

In regards to the issue of deep stops some posts mentioned bubble quantity (BQ) but no one brought up bubble size (BS). Consider these two scenarios:

Small BS and large BQ (VPMB)
Large BS and small BQ (ZHL16C)

VPMB favors deep stops (smaller BS) at the expense of BQ while ZHL16C favors a lower BQ at the expense of larger BS.

With all due respect, I would characterize this as old bubble model dogma generated in the days long before we analysed supersaturation patterns in all tissues over the course of typical bubble and gas content model decompression dives.

Consider a slower tissue that is more supersaturated and for longer late in a bubble model decompression than it would be in a gas content model decompression (I think we can now all agree that this happens). Can you explain to me how the bubble model magically allows a greater number of bubbles to form but somehow controls their size so that they are smaller than would form in the same tissue when less supersaturated during a gas content model decompression?

Simon M

 

[elmo]

Getting back to the wider subject, the early promise of deep stops and bubble models was that they were more efficient. They would get you out of the water in the same or even less time, with less bubbles and greater safety. The concept was embraced with great enthusiasm in the early 2000s based entirely on theoretical attraction and no data. Subsequently, every human study comparing an approach emphasizing deep stops (eg bubble models or ratio deco) has shown the opposite. There is a strong signal in the literature that if you over-emphasize deep stops any advantage of reducing supersaturation in fast tissues early in an ascent is outweighed by the disadvantage of greater supersaturation in slower tissues later after the dive - a construct that I suspect is supported by those results in the Powell study (graphs presented earlier in this thread).

It seems likely that bubble models and decompressions that emphasize deep stops are not the most efficient approach to decompressing. Doesn't mean you should not use them of course, and the material differences in risk compared to other approaches with the same duration may be small. Moreover, I cannot tell you in an evidence based way how far to back away from the bubble model style deep stops that we all embraced in the early 2000s. It is almost certainly not as far as raw Buhlmann, but that information is still to come.

Simon M

Thanks for your response in this thread. Given no comparitive human study has demonstrated benefit in deep stops, I'm curious why you believe the ideal strategy "is almost certainly not as far as raw Buhlmann". Buhlmann's work was based on real dives, and had some rational method to determine the parameters he did. Why do you suspect he could have been neglecting the importance of protecting the fast tissues.

I ask that because I'm interested, not stating I disagree with your suspicion. When I can dive again I can assure you mu gf lo will be lower than my gf hi.

 

[scubadada]

Hi @Dr Simon Mitchell

How do you feel about other markers for decompression stress other than bubbles, the chemokines obtained in the Spisni study for example? Are there other markers that would prove valuable?

 

[dmaziuk]

... Buhlmann's work was based on real dives, and had some rational method to determine the parameters he did. Why do you suspect he could have been neglecting the importance of protecting the fast tissues.

That started long before Buhlmann, in the 1930s I believe. He's been "neglecting the importance" because 5 decades of prior art say he should.

 

[Storker]

Hello,

There has been a lot of discussion in this thread about an important principle exemplified by this question:



Yes, of course it is possible, but it is not particularly useful - not if you are seeking the 'truth in the universe' about optimal decompression strategy.

Lets say you believe in deep stops, and that's why you are interested in 10/85. And lets say that the 10/85 profile for a hypothetical dive prescribes 85 minutes decompression compared to 70 minutes for the 50/85 profile. What would a comparative study of these two profiles tell you?

Could it tell you that one profile is associated with better outcomes than the other? - YES it could.

And lets say your 10/85 profile has better outcomes, does that prove that deep stops are a good idea? - NO it doesn't

It doesn't because the 10/85 profile prescribed a longer decompression and you can't separate that from the effect of the stop depth distribution.

OK, you may say you don't care - all you were interested in was whether your 10/85 profile was better than 50/85, but think about that for a minute. Its not a smart position for an analytical decompression diver to be taking. The question you should be asking yourself is "OK, my favored 10/85 profile says I have to spend 85 minutes decompressing; so am I sure that the distribution of stops prescribed by 10/85 gives me the least risk for THAT length of decompression?"

Let me put it another way. By choosing 10/85 you have effectively chosen to do 85 minutes of decompression. Surely the only question that should now occupy your mind is: "if I'm going to do 85 minutes of decompression, what is the best way to distribute the stops?"

So, lets say you find that 50/70 gives you exactly the same length of decompression as your 10/85 profile, but with slightly less emphasis on deep stops and slightly more emphasis on shallow stops. Surely THAT (or something like it) is the comparison you should be making. Then you can decide whether your stop depth distribution over the duration of decompression you had decided to do is better (or not) than a different distribution of stop depth.

This is the crucial concept of decompression efficiency - the least risk for the same amount of decompression time. The best way to see it is not "what is best VPM vs ZHL or GF this vs GF that", but rather "I've chosen to do X minutes of decompression; what distribution of stops within those X minutes gives me the safest ascent?"

Getting back to the wider subject, the early promise of deep stops and bubble models was that they were more efficient. They would get you out of the water in the same or even less time, with less bubbles and greater safety. The concept was embraced with great enthusiasm in the early 2000s based entirely on theoretical attraction and no data. Subsequently, every human study comparing an approach emphasizing deep stops (eg bubble models or ratio deco) has shown the opposite. There is a strong signal in the literature that if you over-emphasize deep stops any advantage of reducing supersaturation in fast tissues early in an ascent is outweighed by the disadvantage of greater supersaturation in slower tissues later after the dive - a construct that I suspect is supported by those results in the Powell study (graphs presented earlier in this thread).

It seems likely that bubble models and decompressions that emphasize deep stops are not the most efficient approach to decompressing. Doesn't mean you should not use them of course, and the material differences in risk compared to other approaches with the same duration may be small. Moreover, I cannot tell you in an evidence based way how far to back away from the bubble model style deep stops that we all embraced in the early 2000s. It is almost certainly not as far as raw Buhlmann, but that information is still to come.

Simon M

I'm sorry that it isn't possible to double-like a post.

Me, I'd triple-like this post if it were possible.

@Dr Simon Mitchell , thanks for chiming in. I assume you're just as busy doing real work these days as usual. But your commitment to the diving community is really appreciated.

 

[KenGordon]

Buhlmann's work was based on real dives, and had some rational method to determine the parameters he did. Why do you suspect he could have been neglecting the importance of protecting the fast tissues.

in pure Bühlmann ZHL16A the m values are mathematically derived from the tissue half times. This is done with a fairly simple equation. So, while you see a table of M values in the books etc, that table was originally just an equation. In subsequent versions he tweeted the numbers.

Now, it may be that this expresses the longer time slow tissues will be over saturated and so the over saturation limit ought to lower, but I was surprised.

The GF scheme might just be fixing that assumption/simplification a bit more. Other dissolved gas schemes are available.

 

[atdotde]

in pure Bühlmann ZHL16A the m values are mathematically derived from the tissue half times. This is done with a fairly simple equation. So, while you see a table of M values in the books etc, that table was originally just an equation. In subsequent versions he tweeted the numbers.

That is a misconception. The m-values (expressed there as a and b coefficients) were determined empirically and then later he found a fit function for those. But that formula is not a priory (as are the half times)

 

[EFX]

Hi EFX,

With all due respect, I would characterize this as old bubble model dogma generated in the days long before we analysed supersaturation patterns in all tissues over the course of typical bubble and gas content model decompression dives.

Consider a slower tissue that is more supersaturated and for longer late in a bubble model decompression than it would be in a gas content model decompression (I think we can now all agree that this happens). Can you explain to me how the bubble model magically allows a greater number of bubbles to form but somehow controls their size so that they are smaller than would form in the same tissue when less supersaturated during a gas content model decompression?

Simon M

See post #68 and the attached file. Magically!? I guess without any scientific evidence it would seem like magic wouldn't it. I completely agree with you that if supersaturation was all that mattered then stopping shallower rather than deeper is far better. However, I want to know if bubble size (BS) matters also. The only way we're going to know is if we test a deco model that actually calculates BS. It would be nice if we can measure BS.

 

[scubadada]

See post #68 and the attached file. Magically!? I guess without any scientific evidence it would seem like magic wouldn't it. I completely agree with you that if supersaturation was all that mattered then stopping shallower rather than deeper is far better. However, I want to know if bubble size (BS) matters also. The only way we're going to know is if we test a deco model that actually calculates BS. It would be nice if we can measure BS.

So somebody needs to test a "conservative" VPM-B profile with an "aggressive" Buhlmann profile with the same run time Bubble model vs. Gradient Factors redux As an example that could the NEDU profile with VPM-B+4 vs GF 66/66. Something like that would maximize the differences between deep stops with VPM vs. shallow stops with Buhlmann. Somehow, I don't see this happening in the current environment.