Deep Stops Increases DCS

[leadduck]

Incorrect statement because you dont continuously record or measure VGE scores. These are recorded post dive making your statement and conclusion poor at best.
This is also funny coming from someone that used the words "imaginary bubbles" when references bubble models.

VPM's bubbles are imaginary. You cannot expect a correlation between real bubbles in your body and the bubble numbers in VPM. The reason is that VPM's bubble model is only a quite poor and uncalibrated approximation of the mechanisms that cause real bubbles in the human body.

Why do you think there's a problem with measuring VGE post dive vs modeling post dive? VGE tend to peak post-dive while tissues are off-gasing, as there's also a delay between the dive and DCS symptoms.
Having too many bubbles post-dive will also slow down off-gasing of tissues. All of that is important for decompression but missing in VPM. There are better probabilistic models such as Thalmann's LE that avoid a physical bubble model but consider these effects of bubbles and can also predict the time delay quite well.

 

[rossh]

No, the primary and only reliable measure of stress we have are VGE counts and DCS incidence in real dives. If a model, such as VPM, predicts bubble distributions that don't match real world bubbles such as VGE, the reason is simply that the model is poor.
In the case of VPM, I think it's for two reasons: the main mechanism of VGE (other tissues offgasing into blood) is missing in the model (in VPM all tissues are independent and don't off-gas into each other). So you can't expect VPM to predict bubbles in blood.
Secondly, the bubble model parameters haven't been calibrated by experiments. So you can't expect VPM to predict bubbles in other tissues well either.


The main cause for VGE are tissues offgasing into blood; therefore VGE is a very good indicator of overall tissue supersaturation. DCS only show up for high tissue supersaturaion in sensitive tissues; that's the reason why you can predict only a DCS probability from the VGE bubble count.

leadduck wrote: "... the primary and only reliable measure of stress we have are VGE counts ..." "...VGE is a very good indicator of...."

No wrong on all points.

VGE is neither primary or reliable, or accurate.
VGE does not predict DCS.

Refer to Neal Pollock summary slide, and the warnings he explicitly makes - your points above are in breach of these principles.

VPM does not count or pretend to count or monitor VGE. Dissolved models do not do it either. Deco models deal with tissue microbubble growth (extra vascular microbubble growth). VGE is something different.

Don't get tricked by the wishful thinking of some around here to bundle it all together.

 

[leadduck]

One thing that's interesting to note, the original concept of "Pyle Stops" had those deep stops being added to the total bottom time for decompression calculations. So for a 30 minute dive to 150', with Pyle Stops added at 90' and 60', a person would shift to either the 35 minute or 40 minute decompression schedule, depending on how long those stops were. This means that when conducting "Pyle Stops" the original way, you're actually going to be lengthening your shallow stops, not shortening them.

Exactly.
Yount's original VPM paper however set the VPM parameters so that it matches the total runtime of USN tables. Effectively his schedules shortened the shallow stops and moved time to deep stops. Seems it didn't occur to him that he might increase DCS risk by doing so. In his defense, his paper's stated goal was not producing an operational set of diving tables, but it was just a proof of concept whether a physical bubble model could be used to produce diving tables. Others made the mistake of using it for actual diving, misinterpreting both Yount's paper and Pyle's findings.

 

[leadduck]

This is not to pick on VPM, by the way. Any bubble model would do the same.
R..

While I agree with the rest of your posting, I don't with this one.
Deep stops are not an inherent feature of bubble models in general. It is a particular problem of VPM and its descendant RGBM. At the core of the problem lies Yount's simplification of using the same bubble model parameters such as initial critical radius for every compartment. If Yount had set the bubble model parameters differently per compartment, VPM could produce a much shallower curve just as well.

Same thing for Buhlmann's ZHL by the way. If his experiments had shown that protecting fast compartments at the expense of slow ones reduces DCS risk, then the resulting a/b-values would've been different, and his tables would've contained much deeper stops. The reason why ZHL produces shallower ascents is not because it's missing a bubble model, but because the experimentally determined a/b-values are this way.

 

[rossh]

Ross, is your position that bubbles in the blood do not matter, or at most are secondary? That the 'tissue stress' is what matters?

Have experiments been done in animals where bubbles are introduced directly to the blood without the whole hyperbaric exposure thing?

I can't say VGE doesn't matter, because there are clearly many procedural conditions that can be affected by VGE. i.e. the surface deco (Sur-D) is one . Support diver procedures is another dictated by avoiding re-crushing of captured VGE in the lung. The divemaster unhooking the anchor is yet another procedural example that is affected by trapped VGE. And people with the skin rash condition seem to be at a medical disadvantage too.

However, from a purely deco schedule perspective, the decision has been made already - for 45 years we have all collectively ignored VGE in our deco planning, instead used math of tissue pressure stress.

In the last 15 years, since GF and bubble models, and better dive practices have been the focus, we have all collectively improved the track record, and we have the 99.9x% success rate as a result.

Why do want to start looking at VGE now? There is no new problem to solve, we do not have any crisis in injury numbers. Its not going to improve anything.

What the new fuss over VGE really means, is its yet another fake problem to justify more fake answers.

.

 

[rossh]

Exactly.
Yount's original VPM paper however set the VPM parameters so that it matches the total runtime of USN tables. Effectively his schedules shortened the shallow stops and moved time to deep stops. Seems it didn't occur to him that he might increase DCS risk by doing so. In his defense, his paper's stated goal was not producing an operational set of diving tables, but it was just a proof of concept whether a physical bubble model could be used to produce diving tables. Others made the mistake of using it for actual diving, misinterpreting both Yount's paper and Pyle's findings.

VPM calibration was moved further along than just Yount's original work. Erik Maiken did some in his model refinements. He has lots of comparison charts to look at. The work he did in the mid 90's, was further enhanced on the DecoList in the late 90's where modern VPM was formalized. VPM-B (2003) was further calibrated, as the underlying characteristics of VPM-B changed a little from the VPM-A it descended from. You won't much left of the ties between Yount's original calibrations and VPM-B now.

.

 

[Storker]

VPM calibration was moved further along than just Yount's original work. Erik Maiken did some in his model refinements. He has lots of comparison charts to look at. The work he did in the mid 90's, was further enhanced on the DecoList in the late 90's where modern VPM was formalized. VPM-B (2003) was further calibrated, as the underlying characteristics of VPM-B changed a little from the VPM-A it descended from. You won't much left of the ties between Yount's original calibrations and VPM-B now.

.

Which studies and what kind of data was used to calibrate VPM-B?

 

[leadduck]

Why do want to start looking at VGE now? There is no new problem to solve, we do not have any crisis in injury numbers. Its not going to improve anything.

What the new fuss over VGE really means, is its yet another fake problem to justify more fake answers.

There's no new fuss. Doppler ultrasound bubble detection technology is from the 1970s. Measuring VGE is usual in studies, additionally to counting DCS cases. VGE has a high negative predictive value for DCS risk; i.e. low VGE grade means low DCS risk. In the NEDU study, VGE was measured as well. The profile with more DCS cases had also higher VGE grades.

 

[leadduck]

VPM calibration was moved further along than just Yount's original work. Erik Maiken did some in his model refinements. He has lots of comparison charts to look at. The work he did in the mid 90's, was further enhanced on the DecoList in the late 90's where modern VPM was formalized. VPM-B (2003) was further calibrated, as the underlying characteristics of VPM-B changed a little from the VPM-A it descended from. You won't much left of the ties between Yount's original calibrations and VPM-B now.

In Erik Baker's VPM code of 2001,
ftp://decompression.org/Baker/VPM%20Decompression%20Program%20in%20Fortran.pdf
you will find the parameters
Critical_Radius_N2_Microns, Crit_Volume_Parameter_Lambda, Surface_Tension_Gamma, Skin_Compression_GammaC, Regeneration_Time_Constant
to be the same as in Yount's 1986 paper.

When adding Boyle's law adjustment, Baker saw that the VPM-B algorithm became more conservative and suggested to reduce the critical radius and lambda. Great. Tune internal parameters so that the runtime looks OK to the unsuspecting user, but don't worry about parameter calibration by controlled experiments, DCS risk, efficiency or anything.

Which studies and what kind of data was used to calibrate VPM-B?

 

[UWSojourner]

[New Diver] So, what type of decompression model should I use?

[Ross] A bubble model of course, like VPM.

[New Diver] Why do you say that?

[Ross] Because VPM deals directly with micro-bubbles and controls the overall volume of bubbles by the profiles it produces.

[New Diver] And that's important?

[Ross] Absolutely! Too many bubbles can end up giving you a chamber ride. That's why the more sophisticated models like VPM directly control the bubbles. Older models do too, kind of, but they were made before we understood all these sophisticated ideas. And they're all morphing now to look like VPM anyway. So just use the real thing.

[New Diver] Wow, that's cool.

[More experienced diver] I think I read about studies somewhere that said direct measurements of bubbling in divers saw HIGHER levels of bubbles in those divers using bubble models. I think they called them VGE?

[Ross] Yeah, those studies aren't important.

[More experienced diver] Really? Why?

[Ross] Bubble levels in real divers don't matter.

[New diver slowly backing away] Huh? ... Well ...O...K... Good to talk with you ...