A study of decompression sickness using recorded depth-time profiles.

Introduction: 122,129 dives by 10,358 recreational divers were recorded by dive computers from 11 manufacturers in an exploratory study of how dive profile, breathing gas (air or nitrox [N2/O2] mixes), repetitive diving, gender, age, and dive site conditions influenced observed decompression sickness (DCSobs). Thirty-eight reports were judged as DCS. Overall DCSobs was 3.1 cases/10⁴ dives.

Methods: Three dive groups were studied: Basic (live-aboard and shore/dayboat), Cozumel Dive Guides, and Scapa Flow wreck divers. A probabilistic decompression model, BVM(3), controlled dive profile variability. Chi-squared test, t-test, logistic regression, and log-rank tests evaluated statistical associations.

Results: (a) DCSobs was 0.7/10⁴ (Basic), 7.6/10⁴ (Guides), and 17.3/104 (Scapa) and differed after control for dive variability (p ≺ 0.001). (b) DCSobs was greater for 22%-29% nitrox (12.6/10⁴) than for 30%-50% nitrox (2.04/10⁴) (p ≤ 0.0064) which did not differ from air (2.97/1010⁴). (c) For daily repetitive dives (≺12-hour surface intervals (SI)), DCS occurred only following one or two dives (4.3/1010⁴ DCSobs; p ≺ 0.001) where SIs were shorter than after three or more dives. (d) For multiday repetitive dives (SIs ≺ 48 hours), DCS was associated with high multiday repetitive dive counts only for Guides (p = 0.0018). (e) DCSobs decreased with age at 3%/year (p ≤ 0.0144). (f) Males dived deeper (p ≺ 0.001) but for less time than females (p ≺ 0.001).

Conclusion: Collecting dive profiles with dive computers and controlling for profile variability by probabilistic modeling was feasible, but analytical results require independent confirmation due to limited observed DCS. Future studies appear promising if more DCS cases are gathered, stakeholders cooperate, and identified data collection problems are corrected.