O'Dive by Azoth Systems

It remains difficult to completely identify the causes leading to pathogenic bubbles. These causes are often not quantifiable, most of them are interdependent and many – especially those occurring at a microscopic level – remain unknown.

To answer the question of DCS risk prediction, modern scientific techniques use a probabilistic approach involving mathematical models. Those models integrate the quantifiable parameters contributing to the risk – or at least correlated with it – to output the bounds of risk of the accident.

With its expertise in decompression modeling [5], Azoth Systems has developed a series of predictive models called BORA (for Bubble Occurrence – Risk Attrition) to report the results observed in multiple databases. Big databases (defence, offshore) collected in the framework of collaborations and service delivery were analysed during almost 10 years: dive parameters of hundreds of thousands of dives (air, nitrox, trimix, open-circuit and closed-circuit Heliox), the bubble dynamics (Doppler measurements) of several thousand of these dives and the hundreds of accidents they led to, have been studied to calibrate these models [6] [7].

What main results does this data bring to light?

• First of all, they corroborate the previous results and allow to better define their shapes. For the air diving, the risk increases with the depth and the duration of dive and decreases when the total decompression time (stops) increases, according to a severity index Is as represented by the graph below (Figure 3). This result is obtained by the statistical analysis based on logistic regression of three defence databases accumulating more than a million of dives and about 300 cases of accidents (French Navy, US Navy and DRDC).

Figure 3 – The probability of DCS (PDCS) as a function of severity index Is taking into the account the depth, the duration of dive and the total decompression duration. The examples of parameters are presented as dive duration/maximal depth for the square air dives.

• The same trends were proved for deep dives (DRDC database and COMEX heritage), allowing to extend the construction of severity index Is to multi-gas setting, with:
  • A major interest in the introduction of Helium in the breathing gas composition
  • The interest of the oxygen-enriched breathing gas in the decompression phase

     both well characterised with respect to the risk of the accident (logistic regression).

• The pic of vascular microbubbles, observable in precordial or subclavian zones after dive, is the essential data for prediction of risk : accounting for it allows to refine the DCS probability curves (Figure 4). Indeed, the presence of microbubbles acts as an amplificator of risk associated to the dive severity [7]: for dives with depths and durations necessitating decompression stops (i.e. dives on and above the security curve of no decompression) the observed risk ratio between a dive without or with few vascular microbubbles and a dive with a high bubble grade is about 10. This result is based on the analysis of a DRDC database including dive profiles, bubbles measurements in precordial and subclavian zones and accident status (more than 8000 exposures in hyperbaric chamber with more than 100 DCS).

Figure 4 – DCS probability laws as a function of the dive severity index (Is) for 5 levels of post-dive microbubbles (measurements 30’ after surfacing)

• Doppler bubble measurements in subclavian zone characterize the amplification of the risk better than precordial measurements [7]
• The depth of the first stop and the ascent rate to the first stop depend on the type of breathed gas (more or less rich in Helium) and on the exposure severity (COMEX heritage). The introduction of deep stops finds its limits: Azoth cannot evaluate the benefits of deep stops when they are not in this heritage.