The Changing Landscape of Decompression

The Changing Landscape of Decompression

BY DR. R.W. BILL HAMILTON

John S. Haldane

During the last 15 or so years the practice of decompression has changed dramatically, after a long period of what might be considered as steady evolution dating from the original work of Haldane. This is a brief review of that changing landscape. This is not to say that there have not been many creative developments on the way; but to me the thing that is new, the really unique change, is the do-it-yourself capability of divers to generate their own decompression tables. Let us look at a bit of the history.

THIS IS THE WAY WE'VE ALWAYS DONE IT
Building on Haldane's original methods, starting in the 1930s and carrying on through the 1950s, the US Navy developed its decompression tables. This included, of course, the venerable Standard Air Tables with no-stop and repetitive diving capability, and also the technique of surface decompression where the diver leaves the water and finishes the decompression in a chamber. The partial pressure heliox tables were also developed. Commercial diving, both in the US and elsewhere, used the USN tables. Although they existed at the laboratory level, there were no operationally ready tables for use with oxygen-enriched air or O2-N2. In the early 1960s, commercial divers modified the USN heliox tables to eliminate the use of oxygen in the water (it took the Navy another 35 or so years to get around to doing that). This is all there was. In America the prevailing commercial practice was to use surface decompression with oxygen, sur-d/O2. Most commercial companies followed Navy practice, and because of peculiarities in the legal climate they were often subject to devastating litigation when divers got the bends. If the companies deviated from Navy procedures, they were likely to lose in court.

SERIOUS DEVELOPMENT

Dr. Bill Hamilton

DECOMPRESSION IS EMPIRICAL
One fundamental thing about decompression development that is often overlooked is the fact that decompression tables are empirical. That is to say, they are based on experience. The outcome of yesterday's dive leads to tomorrow's table. A rather short feedback loop was used during the development of the US Navy tables, with results for certain time-depth exposures validated by perhaps six man-dives. During commercial deep heliox development, changes in the algorithm were often made based on an exposure of only two divers. Later projects used thousands of dives. The important thing is that experience is a big part of table development. This begs the question of how this experience is incorporated. This leads to the concept of a "decompression model."

What is a "decompression model?" Quantitative scientists in many fields often use mathematical "models" to describe the behavior of, say, a biological system. Such a model might, for example, describe the relationship in numerical terms of the various factors involved in control of the cardiovascular system, incorporating heart rate, exercise level, temperature, oxygen capacity of the blood, and a myriad of other factors that might allow the prediction of blood pressure under given conditions. Some readers might spot the fact that models are not limited to physical and biological science, since they are valuable in other areas such as social science or economics. The point about a model is that it can be used to predict behavior of the system when certain elements are changed. Models, incidentally, have been around a lot longer than computers, but computers have greatly expanded what can be done with them.

Decompression tables are usually generated by some sort of mathematical computation. Drawing on modeling concepts and terminology, many of those who do this sort of thing refer to the computational algorithms used to calculate tables as "models." Although they share many characteristics with true models, it is more realistic to refer to the equations solved in generating a table as something like "computational algorithms" or just "formulas" or "equations." This is not meant to lay to rest the use of the term "model" in decompression. Many models start out trying to imitate some physiological function, but in due course enough changes have to be made that it makes the resulting algorithm more mathematical than physiological, and often more practical than elegant. Experience is incorporated by the judgment of the designers. In fact, an extremely effective set of parameters dedicated to improved air tables were based on the judgment of a group of senior diving medical officers in the Swedish Navy. The values that fit this model were also found, serendipitously, to work for technical dives. In the meantime, at DCIEM in Canada, the Kidd-Stubbs pneumatic analogue device was programmed electronically (thus becoming a model), which yielded the highly regarded DCIEM tables. The Haldane type of computation can be called "deterministic." That is to say, the calculations provide a dive profile for the diver to follow, but without much in the way of a quantitative indication of the reliability, or conversely, the risk, of using such a table. A more systematic method of dealing with the probability of decompression sickness is covered below. With my long-time colleague, Dave Kenyon, I have been calculating tables for 25 or so years, using our program DCAP (Decompression Computation and Analysis Program). We tend to favor a deterministic Haldane-Workman-Schreiner algorithm, but DCAP can do many others. Bob Workman, with the US Navy, showed how to convert Haldane's ratios to M-values, making the algorithm work better for longer and deeper dives. Heinz Schreiner (who was our boss at the Union Carbide-Ocean Systems lab where Dave and I worked) devised a method for handling multiple gases. With adequate dive experience for reference, we learned to make reliable tables (most of the time!).

PROBABILISTIC DECOMPRESSION
Perhaps the ultimate incorporation of the "empirical" concept, and likely the most significant contribution to decompression science since Haldane, is the method of maximum likelihood introduced by Paul Weathersby and colleagues of the US Navy. This allows the developer to analyze a diverse set of dive profiles, to determine their basic probability of decompression sickness, PDCS, and to be able to estimate the PDCS of independent dives using the original data set as a basis. How good the estimate is depends on many things, including how uniform the data set is, how much DCS it includes, and how closely the new dives match the data set. This method was used by the Navy to generate a comprehensive set of air tables with an even more comprehensive repetitive capability; these were intended to have a uniform PDCS throughout, which proved to be difficult. They found that to make the longer and deeper--the more stressful--dives have the same PDCS of around 2%, they had to make their decompression times unreasonably long. They ended up having to live with higher PDCS for these. Even so, because of their complexity (a book 3" thick to replace a few pages in the USN manual), and because some of the shallow dives ended up with shorter allowable no-stop times, the Fleet did not accept the new tables. So the Navy dug out an algorithm developed and tested by Dr. Ed Thalmann at the Navy Experimental Diving Unit with over 3000 man dives. This is the "Mk15/16 Real Time Algorithm" or VVAL18. It has been programmed into a dive computer for the Special Forces, and it will also be used to develop a new set of air tables; it is manufactured by Cochran.

VALIDATION AND THE VALIDATION WORKSHOP
Closely linked to the generation of new tables is the need to validate them. At one time, to be "legal" and have even a slight degree of immunity to unreasonable litigation, this had to be done with laboratory simulations. This is an expensive process at best. To investigate alternatives, the UHMS held a workshop sponsored by NOAA. The Validation Workshop (Schreiner and Hamilton, UHMS, 1989) concluded that tables that were "interpolative" and referenced to known limits could be introduced as provisional under special supervision and circumstances. The judgmental decisions (such as which tables are interpolative, when are they operationally ready, etc.) are the responsibility of the developing organization, which may charge a person or small group with that duty. The Validation Workshop principles have been implemented a number of times.

FLIRTING WITH OXYGEN-ENRICHED AIR
The 1980s saw the introduction to scientific diving of a method of improving decompression by adding extra oxygen to the breathing mix. This had been working well within NOAA's diving program, but it did cause controversy when introduced into recreational diving, for reasons more political than physiological. Little new decompression work was needed for diving with OEA, since one merely finds an established air table and determines the dive depth with the same partial pressure of nitrogen, PN2, as the mix being breathed and uses that table for decompression. One important implication of this is that it started people thinking about using gases other than air for diving.

In my opinion the practice we call technical diving began when Parker Turner and his WKPP colleagues began to add helium to their breathing gas in order to dive to 240fsw or so without excessive narcosis (as would be the case with air). This became possible because special custom decompression tables could be obtained. True, people had been diving with the USN Exceptional Exposure Tables and breathing oxygen in the shallow stops, with good success. This did not eliminate the narcosis of deep air, but it met the definition of "technical diving" in that more than one breathing mix was used on a dive. The British Navy had called diving with rebreathers technical diving for half a century, but the name stuck here.

Left to Right: Bill Gavin, Parker Turner, Bill Main, and Lamar English proudly display the University of Florida's certificates of accomplishment presented by Milledge Murphy after completing the world record Sullivan traverse. Photo ©David Rhea

Working with Parker, we learned to design tables having bottom mixes with nearly optimal oxygen and helium, and to use one or sometimes two OEA mixes as intermediate gases with oxygen at 20 and 10fsw. This is the basic pattern for technical trimix diving that is still in use today.

DEEP STOPS AND LOW SUPERSATURATIONS
Another aspect of this changing landscape is a plethora of new or resurrected computational models. They are dealing with a somewhat anecdotal but important finding that "deeper stops" produce better profiles. Brian Hills tried to tell us about this in the 1970s, along with his "zero supersaturation" approach. Divers have been changing their ascent profiles, often arbitrarily, to include deeper stops. These in many cases have worked well, and there is now what might be called a scramble to come up with models that reflect this and permit the technique to be used universally. Many of the new approaches use one or more hypothetical bubbles (hypothetical just like Haldane's compartments) as the monitor or regulator of the ascent profile. My approach is to watch these efforts and try to take advantage of the ones that work.

CONCLUSION
We are in the midst of a revolution in decompression. The tried and true methods are still working for navies and other major players, but technical divers on the cutting edge are learning how to calculate custom tables and to improve the outcome of their decompressions, and are able in many cases to do this with their own computations. New models are being used and evaluated, and it all seems to be working. This is in keeping with one of my old principles of decompression science: What works, works.