Diving and Hyperbaric Medicine Volume 42 No. 4 December 2012
The incidence of decompression illness in 10 years of scientific
Michael R Dardeau, Neal W Pollock, Christian M McDonald and Michael A Lang
(The incidence of decompression illness in 10 years of scientific diving. Dardeau MR, Pollock NW, McDonald CM and
Lang MA. Diving and Hyperbaric Medicine. 2012;42(4):195-200.)
Background: The American Academy of Underwater Science (AAUS) constitutes the single largest pool of organizations
with scientific diving programmes in North America. Members submit annual summaries of diving activity and any related
Methods: All diving records for a 10-year period between January 1998 and December 2007 were reviewed. Incidents
were independently classified or reclassified by a four-person panel with expertise in scientific diving and diving safety
using a previously published protocol. Subsequent panel discussion produced a single consensus classification of each case.
Results: A total of 95 confirmed incidents were reported in conjunction with 1,019,159 scientific dives, yielding an overall
incidence of 0.93/10,000 person-dives. A total of 33 cases were determined to involve decompression illness (DCI),
encompassing both decompression sickness and air embolism. The incidence of DCI was 0.324/10,000 person-dives,
substantially lower than the rates of 0.9-35.3/10,000 published for recreational, instructional/guided, commercial and/or
Conclusions: Scientific diving safety may be facilitated by a combination of relatively high levels of training and oversight,
the predominance of shallow, no-decompression diving and, possibly, low pressure to complete dives under less than optimal
Decompression illness, decompression sickness, occupational health, safety, scientific diving, epidemiology
Scientific diving is diving performed by individuals
that is necessary to and part of a scientific research or
educational activity, in conjunction with a project or study
under the jurisdiction of any public or private research or
educational institution or similar organization. Divers can
join programmes with or without prior diver training or
experience. Once in the programme they undergo medical
evaluation, skill evaluation and diver training. Diving
operations are required to adhere to formal programme
rules, including depth and, often, task restrictions. Scientific
diver currency requires meeting standards of minimum
diving activity, refresher training and periodic medical
review. A diving officer, acting on behalf of an institutional
diving control board, is typically responsible for the
training and monitoring of all scientific diving activity,
ensuring compliance with rules and safe and effective dive
team operations. While much of the diving is conducted
as shallow, multi-level, no-decompression exposures,
operations are conducted under a range of conditions, from
tropical to polar, fresh and saltwater, sea level and high
altitude, demanding both skill and appropriate real-time
decision-making to prioritize safety. The safety record of
scientific diving programmes is generally recognized as very
good, but the published documentation is limited.
A review of adverse events reported within the scientific
diving community concluded that total pressure-related
injury rates from 1998 to 2005 were similar to those
calculated by the US Occupational Safety and Health
Administration (OSHA) for scientific divers during the late
1970s.1 The computation of injury rate per 100 workers
per year was matched to the earlier OSHA method. Both
studies included minor barotrauma as well as decompression
illness (DCI – the collective term for decompression
sickness [DCS] and arterial gas embolism [AGE]) but did
not address the incidence of only DCI. The limited reports
on the incidence of DCS for scientific diving range from
0/10,000 person-dives in Australia to 2.8/10,000 person-
dives in the Antarctic.2,3 These rates are relatively low when
compared to the 1.4–35.3/10,000 person-dive estimates for
commercial and military diving communities, but additional
documentation is required.3
The American Academy of Underwater Sciences (AAUS)
was formed in 1977 as a collection of organizational
member programmes representing a range of public and
private academic institutions, educational entities and
research units with active involvement in scientific diving.
AAUS membership requires programmes to submit annual
summaries of dives, mode of diving and any incidents
associated with scientific diving. These diving records make
Diving and Hyperbaric Medicine Volume 42 No. 4 December 2012
AAUS a major source of data on scientific diving in North
America. Our goal was to determine the incidence rate of
DCI in a large and diverse record of scientific diving.
We reviewed 10 years of diving records reported by AAUS
organizational members, from January 1998 through
December 2007. Human subjects research approval for
the study was provided by the Divers Alert Network’s
institutional review board.
A four-person review panel, experienced in scientific
diving, the administration of scientific diving programmes,
and diving safety, reviewed all submitted incident reports.
While incident types (hyperbaric, near drowning, etc) were
defined for reporting purposes, there was some latitude for
diving safety officers to determine what was reportable.1 A
four-step filtration process was employed to remove cases
representing other-than-DCI. The first step excluded records
that were ‘non-events’ (submission error) or ‘no injury’
cases. The second excluded cases that were not pressure-
related. The third excluded cases of minor barotrauma (e.g.,
ear squeeze). The remaining cases (possible DCI) were
then classified or reclassified as ‘DCI’, ‘ambiguous’ or
‘not DCI’ using a modified version of previously described
standardized criteria designed for objective post hoc, non-
clinical assessment (Table 1).4 The panel reviewed all
possible DCI cases independently and then came together
to assign final classification based on consensus decision.
Contentious or incompletely documented cases were further
investigated through interviews with involved persons.
Ambiguous cases were considered to be cases of DCI for
the computation of incident rates.
All data in the present study reflect person-dives, that is, even
when a team of two or more dives together, each diver reports
the dive as an individual event. Incident rates and 95%
binomial confidence intervals (CI) are presented as cases
per 10,000 person-exposures. A Chi-square contingency
table was used to compare annual differences in DCI across
reporting years. Significance was accepted at P < 0.05.
Annual scientific diving activity reported by AAUS members
appears in Figure 1. The number of members reporting
increased substantially during the study period (ranging
from 54 to 94), in turn increasing the number of divers
(ranging from 2,716 to 4,101) and person-dives tallied
Cases classified as ‘not DCI’
a) Cases with single dives to less than 9 metres’ sea water (msw) and symptoms that could not be attributed to AGE;
b) Cases with symptom onset times more than 48 hours after the last dive or altitude exposure;
c) Cases with signs and symptoms likely due to a non-diving cause of injury upon review of medical history;
d) Isolated headache, dizziness, anxiety, general weakness, fatigue, or subjective numbness and tingling of both hands and
feet, in the absence of other symptoms or without objective findings.
Note: Cases with no response to recompression were reviewed extensively before being classified as ‘not DCI’.
Cases classified as ‘ambiguous’
a) Cases with sufficient exposure but minimal or atypical symptoms;
b) Cases in which symptoms resolved spontaneously without recompression in less than 20 minutes with surface oxygen
or less than 60 minutes without oxygen;
c) Cases with confounding medical conditions that could explain the symptoms;
d) Combinations of headache, dizziness, anxiety, general weakness, fatigue, and subjective numbness and tingling of both
hands and feet, in the absence of other symptoms or without objective findings.
Cases classified as ‘DCI’
a) Cases with a dive depth of at least 9 msw;
c) Joint/limb pain, skin/lymphatic symptoms, constitutional/non-specific symptoms;
d) Serious neurological, cardiopulmonary, mild neurological and simultaneous presence of pain and constitutional symptoms.
a) Cases with symptom onset in less than 15 minutes post-dive;
b) Cases with cerebral neurological symptoms, signs or findings;
c) Cases with symptom duration greater than 15 minutes.
Note: Rapid ascent, out-of-air incident, or the presence of cardiopulmonary symptoms increased the confidence of an AGE
Case (diagnosis) classification/reclassification (adapted from 4); AGE –arterial gas embolism, DCI – decompression illness
Diving and Hyperbaric Medicine Volume 42 No. 4 December 2012
annually (ranging from 68,598 to 126,831; Figure 1). The
number of dives completed by individual organizational
members varied tremendously, based on the size of the
diving programme and active scientific diving projects.
Most organizational members were American institutions,
with only three to five based outside the United States for
any given year.
The 10-year study period captured 1,019,159 person-dives
and 102 incidents occurring in conjunction with these
exposures. No case involved multiple victims of a single
event. A summary of the case count by year before and
after filtration and/or reclassification to include only DCI
cases appears in Figure 2. Steps to improve the reporting of
organizational member diving activity began with the 2004
reporting cycle and culminated in the implementation of a
formal training programme for new diving safety officers
in 2006. It is possible that the apparent decline in reporting
non-DCI events was associated with heightened awareness
gained through these efforts.
The stepwise filtration of incident reports is summarised in
Table 2. The first step excluded seven cases; five as ‘non-
events’ and two as ‘no injury’. The second step excluded
28 cases as ‘not pressure-related’ (including two fatalities
resulting from medical emergencies – one a myocardial
infarction following an unremarkable checkout dive and the
other a case of unexplained sudden death following a very
short, shallow exposure). The third step excluded 21 cases of
mostly minor barotrauma, yielding 46 cases of possible DCI.
The 46 cases of possible DCI we identified included 13 that
were treated with recompression but then classified as ‘not
DCI’. Of these, five involved a history of back or shoulder
injury that did not respond to hyperbaric treatment, five
involved unrelated medical conditions that were initially
submitted as DCS, and three cases involved symptoms
more likely related to environmental conditions (thermal
stress and anxiety). In only one case did we ‘overturn’ a
physician diagnosis of DCS (in agreement with a follow
up by another physician who ruled out DCS). In one
other case, we retained a classification of DCS when a
physician changed his diagnosis to rule out DCS following
recompression therapy. This incident involved a diver who
reported having shoulder pain pre-dive that felt better at
depth and returned post-dive (he had dived two days earlier).
The pain was fully resolved upon completion of a US Navy
Treatment Table 6. Ultimately, 33 cases were classified as
DCI, 25 with fully evolved symptomology and eight with
ambiguous symptoms. Recompression therapy was reported
to be successful in 28 of the 33 DCI cases; 19 with a single
treatment and nine with multiple treatments.
The 95 valid incident reports yielded an all-events incidence
rate of 0.93/10,000 person-dives. The 33 DCI cases yielded a
rate of DCI of 0.324 per 10,000 person-dives (95% CI 0.234
to 0.424). The annual rates of DCI were not significantly
different (X2 [df 9; crit 16.92] = 3.32), ranging from a low of
0.18/10,000 persons-dives in 2003 to a high of 0.52/10,000
person-dives in 2000 (Figure 2).
The distribution of maximum depth for all captured dives
was 59% < 10 metres’ sea water (msw), 30% 10–18 msw,
9% 19–30 msw and 2% > 30 m (Figure 3). Exposures with a
maximum depth < 10 msw included only one case diagnosed
Number of divers and reported scientific dives by year
All incidents and DCI cases reported
Stepwise case filtration
102 total incidents 5 ‘non-events’, 2 ‘no injury’
95 28 ‘not pressure-related’
67 pressure-related cases 21 minor barotrauma
46 possible DCI cases 13 'not DCI'
33 DCI cases (including 9 ‘ambiguous’ cases)
Diving and Hyperbaric Medicine Volume 42 No. 4 December 2012
as an arterial gas embolism and one ambiguous DCI case.
All but four DCI incidents occurred on dives with maximum
depths between 9 and 30 msw. Two DCS cases occurred on
dives to depths in excess of 40 msw.
DCI is a relatively rare event, requiring monitoring of
exposures over a broad geographic area and a long time
period to yield meaningful rates.5 A number of incidence
measures have been published, but all with much smaller
exposure numbers than the current study (ranging from
14,944 to 700,000 exposures). DAN’s Project Dive
Exploration estimates of the incidence of DCI in the
recreational community to be 2.0–4.0/10,000 person-
dives.4,6,7 This is higher than previously reported rates of
0.90/10,000 person-dives (DCS) and 0.96/10,000 person-
dives (DCI).8,9 DCS rates among divemasters and instructors
have been estimated at 12.7–15.2/10,000 person-dives.10 The
rate of DCI among military sport divers has been estimated
at 1.34/10,000 person-dives.11 Shallow, no-decompression
dives among navy divers produced incidence of DCS of
2.9/10,000 person-dives.12 The US National Oceanic and
Atmospheric Administration (NOAA), which conducts both
working dives as well as scientific dives, reported a DCS
incidence of 1.8/10,000 person-dives.13 The incidence of
DCS in commercial decompression diving has been reported
to be as high as 35.3/10,000 person-dives.14 A more recent
study reported commercial diving DCS incidence rates
ranging from 1.4 to 10.3/10,000 person-dives depending
on the depth of dive operations.15
Long-standing Antarctic scientific diving programmes are
managed by several nations.16,17 The incidence rate of DCS
for Antarctic scientific diving is reported as 2.8/10,000
person-dives (there were no cases of AGE).3 Outside
of Antarctic scientific diving, DCS/DCI incidence rate
estimates in the scientific diving community are lower than
in other diving populations.3 Estimates range from 0/10,000
dives to 0.6/10,000 person-dives.2,18 The zero estimate was
based on 14,944 person-dives and the 0.6/10,000 estimate
included only one case of DCS in 15,711 exposures. The
rate of 0.32 incidents per 10,000 person-dives reported in
the current study falls within this range. The requirements
for routine diving medical surveillance, equipment
maintenance requirements, and additional training and
oversight combined with the predominance of shallow, no-
decompression diving, may result in lower incident rates in
the scientific community than in other diving populations.5,19
Risk estimate efforts have several limitations. A frequent
challenge of epidemiological studies is the accurate
quantification of all relevant activity, effectively the
denominator needed to compute incident rates. Exposure to
DCI among recreational divers has been determined using
prospective studies, or more roughly estimated by surveys
or by surrogate counts such as cylinder fills.4,6,9,11,20 Less
available are the data generated by occupational diving
programmes that require routine logging of both dives and
incidents. An additional complication of studying DCI is
the potential confounding of clustering as injury may be
likely to affect multiple individuals on a shared dive. It is
not always clear whether the reported denominator is the
number of dives or the number of person-dives.18
A major challenge is the sometimes idiosyncratic and often
difficult-to-define nature of DCI. Given the difficulty of
diagnosis, combined with a tendency to treat conservatively,
it is not surprising that many cases treated as DCI may
in retrospect be classified or reclassified as ‘not DCI’
or ‘ambiguous’. Of 435 cases of DCI reported in the
recreational diving community treated with recompression,
85 (20%) were objectively reclassified as not DCI, making it
clear that appropriate incident rate computation is dependent
upon careful evaluation of individual cases.6 In another
study 10 of 104 recompressed cases were reclassified as ‘not
DCI’.11 The current study resulted in the reclassification of
13 of 46 cases from possible DCI to ‘not DCI’. It is important
to note that the classification/reclassification criteria used
in this study were intended to enable objective post hoc
assessment for scientific, not clinical purposes. While the
reviewers in the current study all had extensive experience
with professional diving, none were medical clinicians.
The lack of clinical expertise could lead to errors in case
classification and our findings are not intended to challenge
clinical skill or decision-making.
Notwithstanding evidence that DCS may be over-reported
and often treated conservatively, there are also an unknown
number of unreported cases. Prior to the 1980s, when minor
symptoms of pain were more accepted as a routine part of
diving, divers may have been reluctant to report symptoms.
Even with the current emphasis on early reporting and the
greater accessibility to treatment, some divers may still be
hesitant to report minor symptoms. This situation produces
some uncertainty with the estimated numerator as well.
Depth distribution of DCI cases (count) & dive distribution (%)
Diving and Hyperbaric Medicine Volume 42 No. 4 December 2012
There may also be some variability in incidents deemed
reportable by individual institutions. Some might choose
to report only cases in which time loss or injury occurs.
Others might choose to report all events, regardless of
cost or outcome. Data collection could be improved by
comprehensive definition of reportable events, changing
from annual reporting to near-real-time reporting of
incidents, and adding further structure to case-review
Documenting the degree of risk associated with a given
dive or dives is also problematic. We have presented the
maximum depth of the dive in which the incident occurred
or followed, but this may miss information of potentially
substantial value. Decompression stress can be influenced by
the specific profile of a given dive and also by previous dives
in a series. AAUS diving records do not currently include
depth-time profiles for dives and information regarding
dives preceding an incident dive is frequently incomplete,
effectively making it impossible to quantify decompression
stress independent of outcomes. Our data do confirm a high
level of safety for dives conducted in less than 10 msw depth.
This is certainly expected in terms of decompression safety
and a welcome observation regarding severe barotrauma.
The total number of cases of DCI is too small to make
strong statements regarding the distribution of DCI in the
depth categories greater than 9 msw. The progressive nature
of scientific diver depth authorization does help to ensure
that divers have greater experience for increased working
depths, which may promote safety.
AAUS represents a substantial number of programmes
involved with scientific diving, but many agencies and
organizations conducting scientific diving do not report to
AAUS. For example, NOAA conducted 208,459 person-
dives between 1981 and 2004, of which some would
certainly meet the definition of scientific diving.13 Similarly,
the Alaska Department of Fish and Game made over 10,933
person-dives between 1990 and 2000.21 Ultimately, while
capturing an impressive amount of activity, AAUS dives are
not the only scientific dives performed by US organizations.
Despite the limitations of this study and many others
evaluating diving risk, it does appear that scientific diving
represents one of the safer forms of diving. This safety
may be facilitated by a combination of relatively high
levels of training and oversight, the predominance of
shallow, no-decompression diving and, possibly, low peer
or institutional pressure to complete dives under less than
optimal circumstances. Additional research to compare the
decompression stress of actual exposures, the pressure to
conduct dives, reporting practices, and other variables that
exist between the diving sub-fields could provide useful
insights to understand the real risks.
We reviewed incidents reported in conjunction with 1,019,159
scientific dives documented by AAUS organizational
members from January 1998 through December 2007. A
total of 95 valid incidents were reported, yielding an all-
incidents rate of 0.931/10,000 person-dives. Case-by-case
review indicated that 33 of the cases involved DCI. The
incidence of DCI was 0.324/10,000 person-dives (including
ambiguous cases). This rate is substantially lower than
the previously published rates for recreational diving,
instructional/guide diving, commercial and military diving.
Data collection efforts may be improved by developing
real-time incident reporting guidelines instead of relying
primarily on annual reporting, and developing additional
protocols for immediate follow up of defined cases.
1 Dardeau MR, McDonald CM. Pressure related incidence rates
in scientific diving. In: Pollock NW, Godfrey JM, editors.
Diving for Science 2007. Proceedings of the American
Academy of Underwater Sciences 26th Symposium. Dauphin
Island, AL: American Academy of Underwater Sciences;
2007. p. 111-5.
Carter A, Muller R, Thompson A. The rate of decompression
sickness in scientific diving at the Australian Institute of
Marine Science (Townsville) 1996-2001. SPUMS Journal.
Sayer MDJ, Lang MA, Mercer S. The comparative incidence
of decompression illness in Antarctic scientific divers. In: Lang
MA , Sayer MDJ, editors. Proceedings of the International
Polar Diving Workshop. Washington, DC: Smithsonian
Institution; 2007. p. 191-5.
Vann RD, Freiberger JJ, Caruso JL, Denoble PJ, Pollock NW,
Uguccioni DM, Dovebarger JA. Report on decompression
illness, diving fatalities and Project Dive Exploration.
Durham, NC: Divers Alert Network; 2005.
Sayer M. The international safety record for scientific diving.
SPUMS Journal. 2005;35:117-9.
Vann RD, Denoble PJ, Uguccioni DM, Freiberger JJ, Pollock
NW, Dovenbarger JA, et al. Report on decompression illness,
diving fatalities and Project Dive Exploration. Durham, NC:
Divers Alert Network; 2004.
Pollock NW, Dunford RG, Denoble PJ, Dovenbarger JA,
Caruso JL. Annual diving report - 2008 edition. Durham, NC:
Divers Alert Network; 2008.
Gilliam B. Evaluation of DCS incidence of multi-day repetitive
diving for 77,680 sports dives. In: Lang MA, Vann RD, editors.
Proceedings of repetitive diving workshop. Costa Mesa, CA:
American Academy of Underwater Sciences; 1992. p. 15-24.
Ladd G, Stepan V, Stevens L. New method for counting
recreational scuba dives establishes risk of death and DCI
injury. SPUMS Journal. 2002;30:124-8.
10 Hagberg M, Ornhagen H. Incidence and risk factors for
symptoms of decompression sickness among male and female
dive masters and instructors - a retrospective cohort study.
Undersea Hyperb Med. 2003;30:93-102.
11 Arness MK. Scuba decompression illness and diving fatalities
in an overseas military community. Aviat Space Environ Med.
Diving and Hyperbaric Medicine Volume 42 No. 4 December 2012 Download full-text
12 Flynn ET, Parker EC, Ball R. Risk of decompression sickness
in shallow no-stop diving: an analysis of Naval Safety Center
data 1990-1994. Bethesda, MD: Naval Medical Research
13 Dinsmore DA, Vitch ML. NOAA’s diving accident
management programme: a review of current capabilities
and plans for improvement. In: Godfrey JM, Shumway SE,
editors. Diving for science 2005. Proceedings of the American
Academy of Sciences 24th Symposium. Dauphin Island, AL:
American Academy of Underwater Sciences; 2005. p. 157-69.
14 Imbert JP, Fructus X, Montbarbon S. Short and repetitive
decompressions in air diving procedure: the commercial diving
experience. In: Lang MA, Vann RD, editors. Proceedings
of repetitive diving workshop. Costa Mesa, Ca: American
Academy of Underwater Sciences; 1992. p. 63-72.
15 Luby J. A study of decompression sickness after commercial
air diving in the Northern Arabian Gulf: 1993-95. Occupat
16 Pollock NW. Scientific diving in Antarctica: history and
current practice. Diving Hyperb Med. 2007;37:204-11.
17 Lang MA, Robbins R. Scientific diving under ice: A
40-year bipolar research tool. In: Krupnik I, Lang MA,
Miller SE, editors. Smithsonian at the Poles: Contributions
to International Polar Year Science. Washington, DC:
Smithsonian Institution Scholarly Press; 2009. p. 241-52.
18 Sayer MDJ, Barrington J. Trends in scientific diving: an
analysis of scientific diving operation records, 1970-2004.
Underwater Tech. 2005;26:51-5.
19 Lang MA. The USA scientific diving medical and safety
experience. SPUMS Journal. 2005;35:154-61.
20 Klingmann C, Gonnerman A, Dreyhaupt J,Vent J, Praetorius
M, Plinkert PK. Decompression illness reported in a survey
of 429 recreational divers. Aviat Space Environ Med.
21 Pritchett M. Scientific diving injuries in Southeast Alaska.
In: Jewett SC, editor. Coldwater diving for science 2001.
Proceedings of the American Academy of Sciences 21st
Symposium. Costa Mesa, Ca: American Academy of
Underwater Sciences; 2001. p. 75-7.
Conflict of interest
All four authors are current or past members of the AAUS Board
Submitted: 28 July 2011
Accepted: 25 July 2012
Michael R Dardeau, BS, MS, is Diving Safety Officer at the Dauphin
Island Sea Lab, Dauphin Island, Alabama.
Neal W Pollock, BSc, MPE, PhD, is Research Director at the Divers
Alert Network, Durham, North Carolina.
Christian M McDonald, BS, is Diving Safety Officer at the Scripps
Institution of Oceanography, La Jolla, California.
Michael A Lang, BSc, was Diving Safety Officer at the Smithsonian
Institution, Washington, DC, USA, at the time of submission.
Address for correspondence:
Neal W Pollock
Divers Alert Network
6 West Colony Place
Durham, NC 27705, USA