American Cave Diving Fatalities 1969-2007

International Journal of Aquatic Research and Education
Volume 3 Number 2 Article 7
5-1-2009
American Cave Diving Fatalities 1969-2007

Peter L. Buzzacott
Divers Alert Network, Durham, NC, [email protected]
Erin Zeigler
Divers Alert Network, Durham, NC
Petar Denoble
Richard Vann
Follow this and additional works at: https://scholarworks.bgsu.edu/ijare

How does access to this work benefit you? Let us know!
Recommended Citation
Buzzacott, Peter L.; Zeigler, Erin; Denoble, Petar; and Vann, Richard (2009) "American Cave Diving
Fatalities 1969-2007," International Journal of Aquatic Research and Education: Vol. 3: No. 2, Article 7.
DOI: https://doi.org/10.25035/ijare.03.02.07
Available at: https://scholarworks.bgsu.edu/ijare/vol3/iss2/7
This Research Article is brought to you for free and open access by the Journals at ScholarWorks@BGSU. It has
been accepted for inclusion in International Journal of Aquatic Research and Education by an authorized editor of
ScholarWorks@BGSU.
Buzzacott et al.: American Cave Diving Fatalities 1969-2007
International Journal of Aquatic Research and Education, 2009, 3, 162-177

© 2009 Human Kinetics, Inc.
American Cave Diving Fatalities 1969-2007

Peter L. Buzzacott, Erin Zeigler, Petar Denoble,
and Richard Vann
Fatality records for American cave-diving fatalities (n = 368) occurring between 1969
and 2007 were examined and circumstances preceding each death categorized. Safety
rules breached were noted in each case. The number of deaths per year peaked in the
mid-1970s and has diminished since. Drowning was the most frequent cause of death,
most often after running out of gas, which usually followed getting lost or starting the
dive with insufficient gas. Compared with untrained divers, trained divers tended to
be older, died at deeper depths and further inside caves, carried more cylinders of gas
and more often died alone. Untrained divers were more likely to have dived without a
guideline, without appropriate number of lights, and/or without adequate gas for the
planned dive. Since running out of gas was associated with the greatest number of
fatalities for trained divers, we recommend that gas management rules should receive
the greatest emphasis in cave diving courses.
Cave divers visit environments with inherent hazards that include flowing
water, low or no visibility, a confusing number of possible exits, restrictions,
entanglement in a guideline, and, by definition, a physical overhead barrier pre-
venting divers from ascending to the surface when desperately low on air. The first
recorded scuba dive within a flooded cave occurred in August 1946, described in
detail in the best-seller, The Silent World (Cousteau & Dumas, 1953). The first
cave dives in America using scuba were made in 1952 by the Florida Speleologi-
cal Society (FSS), and one year later, the FSS established the first cave-diving
training program in the United States (Exley, 1994). The first known English-
language text on the exploration of “sumps and vauclusian springs” was a transla-
tion of Caves and Cave Diving by Guy de Lavaur, published in 1957 (DeLavaur,
1957). The first American known to have died while cave diving was Connie Lim-
baugh, who died in 1958 while diving a cave in France, and by 1966 a report on
86 American skin and scuba deaths included 11 deaths in caves or springs (Web-
ster, 1966). This report identified likely contributory causes, such as diving alone
or failing to use a lifeline, suggesting “In virtually all of the accidents, the victims’
disregard of one or more of the recognized rules or procedures for safety was
found to be a contributory cause” (Webster, 1966, p. 5).
The authors are with Divers Alert Network in Durham, North Carolina. Peter Buzzacott is also with
the University of Western Australia, School of Population Health in Crawley. Petar Denoble and
Richard Vann are also with Duke University, Department of Anesthesiology in Durham, NC.
162by ScholarWorks@BGSU, 2009

Published 1
International Journal of Aquatic Research and Education, Vol. 3, No. 2 [2009], Art. 7
Cave Diving Fatalities 163
Cave Diver Training in America

The first American cave diving instructional text was published in 1969, when
Sheck Exley wrote the Dixie Cavern Kings Cave Diving Manual, the same year
the National Association for Cave Diving (NACD) was established by David
Desaultels (Exley, 1994). Today the NACD has over 1,000 current members who
have completed cave diver training (Bauer, 2007). The Cave Diving Manual
appeared in 1972 (Mount, 1972) followed by Safe Cave Diving in 1973 (Mount,
Dickens, & Exley, 1973). This period of American cave diver training material
production was concurrent with a series of diving safety seminars in which annual
fatality reports and possible risk factors were presented (Desautels, 1970, 1972;
Desautels & Briel, 1968). At this time, Desautels added details of 155 known cave
and sinkhole diving fatalities to the Rhode Island University diving fatality data-
base (Schenck, McAniff, Schenck, & Schenck, 1972). The subsequent report
noted the high incidence of multiple fatalities in Florida caves, that cave and sink
divers were dying at deeper depths than other American divers and that the pro-
portion of cases involving untrained divers had actually risen in recent years.
In 1973, Sheck Exley formed and became chairman of the National Speleo-
logical Society (NSS) Cave Diving Section (CDS), to date the largest cave diver
training organization in the world (Exley, 1994). In 1979, Exley published Basic
Cave Diving: A Blueprint for Survival, with descriptions of safe diving practices
and brief case narratives where these practices were not followed and divers died
as a result (Exley, 1979). This close relationship between case review and safe
practice has been repeated in cave diver training materials since. In 1991, the
method was adapted to examine diving practices at shipwrecks (Chowdhury,
1991), where human-created overhead environments present challenges and haz-
ards that are similar to cave diving. That same year, Martyn Farr published a case
series of cave diving fatalities with potential root causes as an appendix (Farr,
2003). Today, many cave-diver training texts review cave-diving fatalities as part
of the cave-diving course (Harrison, Bulling, & Richardson, 1998; Wohlers,
1993). The NSS-CDS continued to produce American cave diver training texts
throughout the 1980s. Other training agencies followed suit in the 1990s and grad-
ually as landowners became aware of the availability of training, untrained divers
found it increasingly difficult to access popular caves (Farr, 2003).
When Blueprint for Survival was published in 1979, cave diver training agen-
cies accepted that most cave diving accidents involved three initial common risk
factors (Zumrick, Prosser, & Grey, 1988):
• Failing to dive with a continuous guideline to the surface;
• Failing to limit the dive’s penetration to 1/3 of the starting air volume; and/or
• Diving beyond a safe depth for the type of gas being used.
In 1984, two additional rules were added: that all divers entering caves should
have cave diving training, and each diver should have three sources of light, each
capable of outlasting the dive (Jablonski, 2001). These five rules now form the
core of many cave diver courses (Bozanic, 1997; Harrison et al., 1998; Jablonski,
2001).
https://scholarworks.bgsu.edu/ijare/vol3/iss2/7
DOI: https://doi.org/10.25035/ijare.03.02.07 2
164 Buzzacott et al.
Cave Diving Fatality Reviews

A case series of 11 British diving fatalities was published in 1980 (Churcher &
Lloyd, 1980) in which overrunning the air margin was implicated in eight of those
deaths. This review was soon followed by a descriptive case series of 11 cave and
sinkhole diving fatalities occurring in South Australia between 1950 and 1980,
published in 1981 (Horne, 1981) and revised in 1987 to include a further three
cases (Horne, 1987). Horne found the majority of cave-diving fatalities (91%)
involved running out of air as a result of either deep-diving with inadequate gas
reserves or getting lost in silty, low visibility conditions. Horne concluded “ . . . it
is absolutely essential for all prospective cave divers to learn about the special
hazards and techniques involved in the activity, before they enter water filled
caves” (Horne, 1987, p. 64). Australian cave diving fatalities were included in
annual diving fatality reports between 1972 and 1998 and in compendiums of
these reports (Walker, 1998, 2002). Interestingly, cave diving deaths virtually
ceased altogether in Australia after the Cave Divers Association of Australia was
formed in 1973, following a number of high profile cave diving deaths (Bulling &
Harrison, 1998) and access to popular caves became restricted to trained cave
divers only (Farr, 2003).
Since 1992 through the present day, the NSS have published case narratives
of recent cave diving fatalities, in the hope that future cave divers will not repeat
contributory errors of judgment (Grey, 1992). In 1997, an analysis of 287 cave
diving fatalities that had occurred within Florida was conducted and, where infor-
mation was sufficient to suggest likely circumstances preceding each case, poten-
tial causal factors were proposed (Byrd & Hamilton, 1997). The authors noted
that drowning accounted for the majority of causes of death (93%) and that a lack
of cave diver training was the leading contributing factor, followed by the failure
to use a continuous guideline back to safety.
In the United Kingdom, cave diving fatalities are included within the British
Cave Rescue Council (BCRC) annual incident reports (Allwright, 2006) and also
within the British Sub-Aqua Club’s annual incident reports (Cumming, 2005,
2006). In 1999, the BCRC published an analysis of British cave rescues (All-
wright, 1999), including seven cave diving incidents, modeled on an earlier analy-
sis of 25 French cave diving fatalities (Dodelin, 1997).
In 2005, the NSS published a summary and analysis of ten American cave
diving fatalities (Gooking, Green, & Burns, 2005), identifying likely contributing
factors. In 2006, a British analysis (Brock, 2006) compared nine cave diving fatal-
ities occurring within the previous 25 years to 11 that had occurred in the 34 years
before that period (Churcher & Lloyd, 1980). The latter report concluded, “the
most significant hazard to experienced divers is inadequate line management and
the most frequent major hazard to inexperienced divers is lack of training”
(Churcher & Lloyd, 1980, p. 1).
The work summarized above identified events that may have contributed to
cave diving deaths and led to training programs with rules to make these events
less likely. The effectiveness of these programs is suggested by a decline in the
proportion of deaths involving untrained divers. We reviewed 368 cave diving
deaths to identify the rules most frequently violated by either trained or untrained
divers. We hypothesize that different circumstances apply to fatalities in these
groups and argue that for trained divers, the most important rules are associated
Published by ScholarWorks@BGSU, 2009 3
with the most frequent incidence of deaths. We further propose that greater empha-
sis given these rules during training might make cave diving even safer with fewer
cave diving fatalities.
Method
Data Sources
American diving fatalities were investigated by John McAniff of the University of
Rhode Island between 1970 and 1989 (McAniff, 1970, 1989) in conjunction with
the Divers Alert Network (DAN) from 1989 until his retirement in 1995 (Doven-
barger, 1991) and by DAN since 1995 through the present day (Vann et al., 2005).
Cave diving fatality records contained within the collection at DAN were reviewed
and the circumstances surrounding each death were classified according to a
system similar in practice to one used in 1997 to classify factors relating to cave
diving fatalities in Florida alone (Byrd & Hamilton, 1997) and to the method
employed when classifying 103 shipwreck diving fatalities in 1992 (Chowdhury,
1992). Up to four stages in the chain of events leading up to each case were clas-
sified using a process of root cause analysis. Root cause analysis is defined by the
USA Department of Veteran’s Affairs as “a process for identifying the basic or
contributing causal factors that underlie variations in performance associated with
adverse events or close calls”(NCPS, 2008). The four stages were the cause of
death, the preceding event immediately before the death, the harmful action
known to have occurred immediately before the preceding event, and the event
that led to, or triggered, the harmful action.
Causes of death were taken from autopsy, medical examiner reports, or death
certificates. Circumstances preceding each death were classified according to evi-
dence contained within each file, such as witness statements, police reports, equip-
ment examinations, gas analysis reports, and other documents. As each new clas-
sification was identified and added to the instrument, the sequential number of
cases reviewed was recorded and growth of the instrument was charted (see Figure
1).
We also noted cave diving safety rules that were broken in each case and then,
when a broken rule directly affected the chain of events, breaking that rule was
classed as relevant. Breaking the training rule was deemed relevant only when a
diver was known to be untrained and known to have broken at least one other rule
classed as relevant to the death. Merely being untrained was not deemed relevant
if no other rules were broken and classed as relevant.
Each case was then again reviewed and the chain of events classified by a
second reviewer. The two initial classifications were compared and overall inter-
rater agreement calculated for up to four stages classified by the reviewers. Exact
or proportion of agreement was measured for rules classed as broken or relevant.
Disparate classifications were reexamined and the disparity resolved between
reviewers. Raw initial agreement between raters during the first 316 cases mea-
sured just 52% when classifying which of the five rules had been broken (344/666
rules classed as broken), and 61% when classifying which of any broken rules had
been directly relevant to the death (211/348 rules classed as broken and
relevant).
Figure 1 — Cases reviewed and growth of the instrument.

Published 5
Figure 1 (continued)
DOI: https://doi.org/10.25035/ijare.03.02.07 167 6
Statistics
Data were managed using Microsoft Access and analyzed using SAS version 9.1
(SAS Inc, Cary, NC). Interrater agreement was measured using simple kappa,
proportion of agreement adjusted for chance occurrence. Scores between 0.41–
0.60 indicate fair agreement, 0.61–0.80 good agreement, 0.81–0.90 very good
agreement, and between 0.91 and 1.00 excellent agreement (Byrt, 1996). Differ-
ences in means were tested using Student’s T-tests for normally distributed con-
tinuous variables and by Wilcoxon Rank-Sum test for data with skewed distribu-
tions (depth and distance inside cave). Odds ratios are reported with 95%
confidence intervals. Variables of interest were then fitted to a logistic model.
Nonsignificant variables (p ≤ .05) were removed by backward elimination.
Results
Initially, the classification instrument started with just two possible causes of
death (drowning and unknown), three possible preceding events, eight possible
harmful acts and 14 possible triggers (the fourth tier). These were selected as a
starting point by reviewing 50 pilot cases before the commencement of the clas-
sification phase. The subsequent growth of the instrument over the review of the
first 373 cases is shown in Figure 2. The finished instrument contained five causes
of death, seven preceding events, 21 harmful acts, and 25 triggers. The last addi-
tion to the instrument occurred while reviewing the 244th case, and there were no
further categories needed to classify the remaining 129 cases. Of the 373 cases
co-reviewed, 316 occurred between 1969 and 2007 inclusive, and of the last 129
Figure 2 — Growth of classification categories by number of cases reviewed.

cases reviewed by both reviewers, 127 occurred during this same time period,
which is when cave-diver training was available, as described earlier. For each of
the four stages classified, interrater agreements are presented in Table 1. Measure-
ment of initial agreement was completed with case number 316 and the remaining
52 cases were reviewed by single reviewers. As indicated by the last row of Table
1, at that stage initial agreement was generally good.
At the completion of the study, the final dataset comprised 368 divers who
died while diving in caves between 1969 and 2007. Males accounted for 95% (n
= 350) of all fatalities. Of the 368 cave diving fatalities, 329 (90%) died in caves
within the USA, 14 (4%) in Mexico, and 9 (2%) in the Bahamas. The remaining
16 divers (4%) were distributed between 10 other countries. Most of the deaths (n
= 287) within America occurred in Florida (87%), 6 (2%) died in Texas, 6 (2%) in
Missouri, 4 (1%) in Georgia, 4 (1%) in Hawaii, and the remaining 22 divers (7%)
were distributed across 16 other states. In Florida, three caves accounted for 111
of the 287 deaths (39%). They were Peacock Slough (n = 46, 16%), the Devil’s
System (n = 34, 12%), and Little River Spring (n = 31, 11%).
Of the 275 divers whose occupation was known, the two most common were
student (n = 89, 32%) and U.S. Military (n = 34, 12%). Professional divers and/or
diving instructors accounted for 15 of the 275 (5%). Of 199 divers for whom the
level of recreational diving certification was known, 26 (13%) were rated diving
instructors. Marital status was known for 262 divers; 96 were single (37%), 154
married (59%), and 12 were divorced/widowed (5%).
Seventy-four divers (20%) were known to have been trained in cave-diving
techniques, while 208 (57%) were reported to have not completed any cave diving
training. The training status of the remaining 86 (23%) divers could not be deter-
mined from information contained within the records. The number of deaths per
year within the dataset appears to diminish over time, as shown in Figure 3, while
the proportion of divers who were trained appears to have increased over the same
period, as shown in Figure 4.
As shown in Table 2 binary tests indicated that compared with untrained
divers, trained divers tended to be older (p < .01), to die at deeper depths (p < .01),
further into the cave (p < .01), to carry more cylinders (p < .01), and to die alone
(p < .01). Table 2 also illustrates the proportion of deaths where divers were known
to have broken any of the five rules and when breaking those rules directly con-
tributed to each death.
Table 1 Measures of Interrater Agreement

Cause of Preceding Harmful
Death Event Action Trigger
Proportion of agreement (n 94% 91% 72% 89%
= 316)
Proportion of agreement (last 90% 98% 90% 93%
127)
kappa (n = 316) 0.77 0.61 0.32 0.65
kappa (last 127) 0.76 0.88 0.68 0.57
Figure 3 — Number of reported cave diving deaths by year.
Figure 4 — Proportion of cave diving deaths by trained divers, by year.
In addition to breaking the training rule, untrained divers were significantly

more likely than trained divers to have broken the continuous guideline rule (OR
10.9, 95% CI 5.1, 23.0), the three lights rule (OR 15.6, 95% CI 6.8, 35.7), and the
thirds rule (OR 2.8, 95% CI 1.6, 5.1). Among untrained divers, it was significantly
more likely that breaking the rule of thirds would be relevant (OR 2.0, 95% CI 1.1,
3.7) and that breaking the continuous guideline rule would be relevant (OR 6.9,

Table 2 Characteristics of Cave Divers by Training Status

Trained (n = 74) Untrained (n = 208) Overall (n = 368)
Age in years ū (SD, n) 37.5 (11.4, 72) 26.7 (8.9, 205) 28.8 (10.3, 363)
Depth in feet ū (SD, n) 155 (127, 49) 74 (37, 170) 91 (71, 268)
Distance in feet ū (SD, n) 750 (817, 41) 258 (406, 133) 374 (544, 210)
Number of cylinders n (%)
Single cylinder 5 (7) 143 (69) 181 (49)
More than one cylinder 37 (50) 17 (8) 65 (18)
Total 42 (57) 160 (77) 246 (67)
Single or multiple n (%)
Single deaths 58 (78) 89 (43) 188 (51)
Multiple deaths 16 (22) 119 (57) 180 (49)
Rules Broken n (%)
Thirds Rule 19 (26)a 103 (50)a 152 (41)a
Training Rule 0 (0)a 208 (100)a 208 (57)a
Line Rule 9 (12)a 125 (60)a 162 (44)a
Gas Rule 12 (16)a 6 (3)a 25 (6.8)a
Lights Rule 7 (9)a 129 (62)a 167 (45)a
Broken and Relevant n (%)
Thirds Rule 16 (84%)b 74 (72%)b 109 (72%)b
Training Rule — 127 (61%)b 127 (61%)b
Line Rule 5 (56%)b 69 (55%)b 87 (54%)b
Gas Rule 3 (25%)b 2 (33%)b 8 (32%)b
Lights Rule 0 (0%)b 2 (2%)b 3 (2%)b
Note. ū = Mean, SD = standard deviation, n = number of cases. a = percentage of divers who broke rule, by training
status; b = percentage of times broken rule deemed relevant, by training status.
95% CI 2.6, 17.8). Trained divers were more likely to dive with a gas that was
inappropriate for the depth, such as deep-air diving (OR 6.5, 95% CI 2.3, 18.2),
but breaking that rule was not more likely to be relevant.
Training did not appear to significantly change the likely cause of death.
Causes of death among trained and untrained divers are presented in Table 3. For
those divers who drowned (n = 294), the event that immediately preceded death is
presented in Table 4. Untrained divers were more likely to have run out of air
before drowning (OR 3.3, 95% CI 1.7, 6.3). For those divers who ran out of air (n
= 211), the harmful actions that immediately preceded running out of air are
shown in Table 5. Getting lost or making a dive with insufficient air preceded 76%
of divers running out of air. Circumstances that preceded getting lost (n = 93) are
presented in Table 6 and circumstances preceding insufficient gas (n = 67) are
presented in Table 7.
Fitting the main effects to a logistic model, and removing least significant
effects by backward elimination, the remaining differences between trained and
untrained divers were that older divers who died in caves were more likely cave
Table 3 Cause of Death by Training Status

Cause of Death Trained (n = 74) Untrained (n = 208) Overall (n = 368)
Drowning 47 (64%) 180 (87%) 294 (80%)
AGE 7 (9%) 1 (0%) 11 (3%)
Cardiac 3 (4%) 1 (0%) 4 (1%)
Anoxia 1 (1%) 1 (0%) 2 (1%)
Unknown 16 (22%) 25 (12%) 57 (15%)
Table 4 Preceding Events by Training Status for Drowning

Preceding Event Trained (n = 47) Untrained (n = 180) Overall (n = 294)
Ran out of air 31 (66%) 134 (74%) 211 (72%)
Did not run out of air 15 (32%) 31 (17%) 56 (19%)
Unknown 1 (2%) 15 (8%) 27 (9%)
Table 5 Harmful Actions Preceding Air Depletion by Training Status

Harmful Action Trained (n = 31) Untrained (n = 134) Overall (n = 211)
Lost 14 (45%) 62 (46%) 93 (44%)
Entangled 5 (16%) 9 (7%) 20 (9%)
Stuck in restriction 1 (3%) 3 (2%) 4 (2%)
Entrapped 1 (3%) 1 (1%) 2 (1%)
Equipment failure 1 (3%) 3 (2%) 7 (3%)
Insufficient gas 8 (26%) 49 (37%) 67 (32%)
Unknown 1 (3%) 7 (5%) 18 (9%)
Table 6 Circumstances Preceding Getting Lost

Preceding Getting
Lost Trained (n = 14) Untrained (n = 62) Overall (n = 93)
Silt-out 5 (36%) 33 (53%) 41 (44%)
No line 4 (29%) 27 (44%) 44 (47%)
Wrong turn 4 (29%) 2 (3%) 7 (8%)
Unknown 1 (7%) 0 (0%) 1 (1%)
Table 7 Circumstances That Preceded Diving With Insufficient Gas

Preceding Insufficient
Gas Trained (n = 8) Untrained (n = 49) Overall (n = 67)
Poor gas planning 8 (100%) 48 (98%) 65 (97%)
Current 0 (0%) 1 (2%) 2 (3%)
Unknown 0 (0%) 0 (0%) 0 (0%)

Published 11
diving trained (OR 1.08 per year, 95% CI 1.05, 1.12) and that cave diving trained
divers were less likely to have died in a manner where breaking the continuous
line rule was relevant (OR 8.99, 95% CI 4.00, 20.20) and more likely to have died
alone, rather than in a multiple fatality (OR 3.49, 95% CI 1.69, 7.18). Dividing the
dataset chronologically into two equal halves, early and late, and running a similar
regression analysis, the differences between early and late divers were almost
identical to the differences between trained and untrained cave divers.
Discussion
Fatality investigations aim to improve practice through analysis of the chain of
events preceding each fatality. Causality is implied in order for a process of root
cause analysis to yield results with bearing on practice. Temporality alone, how-
ever, does not prove causality and the results of this study should be interpreted
with this in mind. Nevertheless, the final version of the classification instrument
attached as an appendix when used with defined inclusion criteria, produced sub-
stantial agreement between raters. Of course, rater agreement diminished the fur-
ther back the reviewers tried to peer. The authors intend for this instrument to be
used to classify other diving fatality datasets, thus enabling direct comparison(s).
This is the first such instrument developed for diving fatality classification with
established interrater objectivity and, as such, it represents an advancement in the
field of diving fatality investigation.
It is encouraging that fewer cave diving deaths are recorded in the U.S. each
year. Without knowledge of the total number of participants or number of dives
made per year, one cannot say whether the overall situation has improved in recent
years, other than to say recently fewer cave diving deaths are recorded annually by
DAN than has been recorded in previous years. From an epidemiological point of
view, these data are far from conclusive. From a public health perspective, how-
ever, decisions regarding public safety are often made with less than complete
information, and we are confident this study adds significantly to knowledge of
cave diving fatalities.
With tougher access conditions, negative publicity following cave diving
deaths and/or the rising expense of cave diver training and equipment, it could be
argued that fewer divers are diving caves than were in previous years. Fewer
untrained divers are reported to be dying in caves in recent years, and the three
caves accounting for the most deaths now each have access restricted to trained
divers only, so it is possible these access restrictions could be, at least in part,
responsible for the drop in number of annual cave diving fatalities. This would be
in keeping with the Australian experience, as described earlier in the background.
So too, specialized equipment has become more available over time. This study
does not claim to prove that existing cave diver training is responsible for the
decline in number of cave-diving deaths recorded annually in the U.S. Improved
equipment and restricted access might equally explain why the differences
between trained and untrained divers are the same as the differences between
early and late divers.
The most common cause of death in flooded caves remains drowning, itself a
diagnosis of exclusion and of little use to the design of safety initiatives. Immedi-
ately before drowning, the majority of divers ran out of air, and this was more
likely among untrained divers. We observed that 76% of divers who ran out of air
did so after either getting lost or starting the dive with insufficient gas for the
planned excursion. How to avoid getting lost and how to plan one’s gas needs are
covered by most, if not all, cave diver training courses. This could explain, at least
in part, why untrained divers were more likely to have run out of air.
Of the 67 divers known to have run out of gas after poorly planning their gas
needs, only two were known to have had insufficient gas for dives in the strong
currents they encountered. The remaining 65 (97%) should have had sufficient gas
to escape the cave if they had turned back at one-third of their gas usage. In this
sample at least, training appears to have made no difference to the consequence of
poor planning. Put simply, regardless of training status, failing to plan adequately
for one’s gas needs leads to running out of air and drowning.
Training does appear to be associated with the circumstances that preceded
getting lost. Untrained divers were significantly more likely to have dived without
a guideline, thus breaking one of the five “golden rules” and were also more likely
to have gotten lost after a silt-out was recorded. Trained divers were more likely
to have taken a “wrong turn,” perhaps a consequence of the increasing complexity
of long cave systems, now often with multiple guidelines and marked routes.
Raters initially agreed which rules were broken only half of the time (51%),
which illustrates how difficult it is to find obvious evidence of rule breaking from
data contained within the files. Half of the classifications needed a third, joint
review. The importance of each of the five rules was implied by the proportion of
times breaking them was relevant to the outcome. The “gas-thirds” rule was the
most likely to become relevant when broken, followed by the failure to have
received training rule. Trained divers broke the continuous guideline rule far less
often than untrained divers (12% vs. 60%), although when they did it became
relevant just as often (56% of the time), which should reinforce the importance of
this rule to all who dive in caves. Using the wrong blend of gas for the depth, for
example diving with air deeper than 40 m, did not commonly precede death.
Rarely was there evidence that taking fewer than the recommended three lights
contributed to a death. That is not to say breaking the lights rule did not more
often play a part in the outcome, but merely that the evidence did not clearly show
that carrying an additional light would have prevented a diver from running out of
air, the most common event to precede drowning.
Worldwide, both diving and nondiving populations alike are known to be
growing older, and, as the proportion of trained divers among the dead has risen
over time, it is no surprise then the mean age for trained divers (37.5 years) would
be greater than for untrained divers (26.7 years).
Conclusions
Table 2 suggested that the fatal exposures were greater for trained than untrained
divers as indicated by depth (155 feet trained, 74 feet untrained), distance traveled
(750 feet trained, 258 feet untrained), cylinders carried (50% more than one cyl-
inder for trained, 69% only one cylinder for untrained), and diving alone (78%
trained, 43% untrained). When a broken rule was relevant, the most critical broken
rules were the thirds (84% trained, 72% untrained) and line rules (56% trained,
55% untrained). Being untrained was relevant for 61% of untrained divers.

The most important preceding event was running out of air (66% trained,
74% untrained), while the most important harmful actions were becoming lost
(45% trained, 46% untrained) and having insufficient gas (26% trained, 37%
untrained). The circumstances that preceded getting lost were silt-out (36%
trained, 53% untrained), no use of a guideline (29% trained, 44% untrained), and
wrong turn (29% trained, 3% untrained). The circumstance that most often pre-
ceded a dive with insufficient gas was poor gas planning (100% trained, 98%
untrained).
Regarding untrained divers, one need conclude no more than they ought not
dive in caves. For cave trained divers, however, there might be fewer fatalities if
training placed greater emphasis on closely observing all five rules, especially the
rule of thirds and using a continuous guideline. Engaging in deeper and longer
dives as well as diving alone also were strongly represented among trained divers,
but dives of this nature represent personal choice and are not covered by rules. We
suggest cave divers remember the Pareto principle (Juran, 1945; Steindl, 2008):
the majority of accidents are associated with relatively few potential causes. In
this study of cave diving fatalities, the potential causes preceding death were a
failure to use a continuous guideline to the surface and bringing inadequate gas
for the intended dive. We recommend every cave diver give additional consider-
ation to the use of a guideline and turn around after using one third of the gas rules
above all others.
References
Allwright, P. (1999). An analysis of British cave rescues for the period 1989 to 1998. Sed-
bergh, Cumbria: British Cave Rescue Council.
Allwright, P. (2006). Incident Report for Period 1st January 2005 - 31st December 2005.
Sedbergh, Cumbria: British Cave Rescue Council
Bauer, J. (2007). Minutes 3rd Quarter Board of Directors Meeting. NACD Journal, 39(4th
Qtr), 35-37.
Bozanic, J.E. (1997). AAUS standards for scientific diving operations in cave and cavern
environments: A proposal. Melbourne, FL: Island Caves Research Centre.
Brock, D. (2006). The learning curve: A quantitative analysis of fatal British cave diving
incidents from 1980 to 2005. CDG Newsletter(159), 7.
Bulling, G., & Harrison, G. (1998). Cavern sinkhole diver course student notes. Mount
Gambier, SA: Cave Divers Association of Australia.
Byrd, J.H., & Hamilton, W.F. (1997). Underwater cave diving fatalities in Florida: A review
and analysis. Journal of Forensic Sciences, 42(5), 807–811.
Byrt, T. (1996). How good is that agreement? (Letter to the editor). Epidemiology (Cam-
bridge, Mass.), 7, 561.
Chowdhury, B. (1991). Wreck diving accident analysis 1970-1990 (No. URI-SRR-91-22).
Kingston, RI: National Underwater Accident Data Centre (NUADC).
Chowdhury, B. (1992, 12th October 1992). Training and diving lessons from wreck diving
accident analysis 1970-1990. Paper presented at the Accident Prevention: A Commu-
nity Workshop, Philadelphia, Pennsylvania.
Churcher, R.A., & Lloyd, O.C. (1980). British cave diving accidents. Proc. University of
Bristol. Spelaeological Society, 15(3), 161–182.
Cousteau, J., & Dumas, F. (1953). The silent world. London, UK: Hamish Hamilton Ltd.
Cumming, B. (2005). NDC diving incidents report. South Wirrel, Cheshire: The British
Sub-Aqua Club.
Cumming, B. (2006). NDC diving incidents report. South Wirral, Cheshire: British Sub-
Aqua Club.
DeLavaur, G. (1957). Caves and cave diving. London: Robert Hale.
Desautels, D.A. (1970). A ten year survey of skin and scuba diving fatalities in the state of
Florida. High Springs, FL: National Association for Cave Diving (pp. 10).
Desautels, D.A. (1972). Cave diving drowning statistics. A twelve year study. Paper pre-
sented at the National Association for Cave Diving, FL.
Desautels, D.A., & Briel, L. (1968). Diving deaths in Florida: A survey of vital statis-
tics. 1960-1967. Paper presented at the First annual seminar on cave diving, Camp
Kalaqua, Hornsby Springs, FL.
Dodelin, C. (1997). Accidents speleologiques en France de 1985 a 1995. INFO-Speleo
Secours Francais(45), 12-16.
Dovenbarger, J. (1991). Report on 1989 diving accidents and fatalities. Durham, NC:
Divers Alert Network.
Exley, S. (1979). Basic cave diving: A blueprint for survival. Brandford, FL: National
Speleological Society - Cave Diving Section.
Exley, S. (1994). Caverns measureless to man. St. Louis, MO: Cave Books.
Farr, M. (2003). The darkness beckons (Rev. ed.). London: Diadem Books.
Gooking, J., Green, K., & Burns, J. (2005). American caving accidents. NSS News, 63(8
part 2), 48.
Grey, H.V. (1992). Recent accidents: Editorial. Underwater Speleology, 19(5), 8-9.
Harrison, G., Bulling, G., & Richardson, T. (1998). Cave diver course student notes. Mount
Gambier, SA: Cave Divers Association of Australia.
Horne, P. (1981). South Australian diving fatalities 1950 - 1980. Hove, SA: P. Horne.
Horne, P. (1987). South Australian diving fatalities 1950 - 1985 (Rev. ed). Adelaide, SA:
P. Horne.
Jablonski, J. (2001). Beyond the daylight zone: The fundamentals of cave diving. High
Springs, FL: Global Underwater Explorers.
Juran, J. (1945). Management of inspection and quality control. New York, London: Harper
and Bros.
McAniff, J. (1970). U.S. underwater diving fatality statistics, 1970. Kingston, RI: National
Underwater Accident Data Center.
McAniff, J. (1989). U.S. underwater diving fatality statistics, 1988: With a preliminary
assessment of 1989 fatalities. Kingston, RI: National Underwater Accident Data
Center.
Mount, T. (1972). Cave diving manual. High Springs, FL: National Association for Cave
Diving.
Mount, T. Dickens, R., & Exley, S. (1973). Safe cave diving. High Springs, FL: National
Association for Cave Diving.
NCPS. (2008). Glossary of patient safety terms: root cause analysis. Retrieved September
23, 2008, from http://www.va.gov/ncps/glossary.html
Schenck, H.V., McAniff, J., Schenck, M.L., & Schenck, H.V. (1972). Diving accident
survey, 1946-1970, including 503 known fatalities (No. 5). Kingston, RI: University
of Rhode Island.
Steindl, J. (2008). Pareto distribution. In S.N. Durlauf & L.E. Blume (Eds.), The new
palgrave dictionary of economics. 2nd ed. Houndmills Basingstoke: Palgrave Mac-
millan.
Vann, R.D., Denoble, P., Dovenbarger, J., Freiberger, J.J., Pollock, N.W., Caruso, J.L., et
al. (2005). Report on decompression illness, diving fatalities and project dive explora-
tion: 2005 edition (Based on 2003 data). Durham, NC: Divers Alert Network.
Walker, D. (1998). Report on Australian diving deaths 1972-1993. Melbourne, FL: J.L.
Publications.

Walker, D. (2002). Report on Australian diving deaths 1994-1998. Ashburton, Victoria:

Divers Alert Network (DAN) S.E. Asia-Pacific Limited.
Webster, D.P. (1966). Skin and scuba diving fatalities in the United States. Public Health
Reports, 81, 703–711.
Wohlers, B. (1993). Cavern diver specialty course instructor outline. Santa Ana, CA: Inter-
national PADI.
Zumrick, J.L., Prosser, J.J., & Grey, H.V. (1988). NSS cavern diving manual (1st ed.).
Branford, FL: National Speleological Society Cave Diving Section.

American Cave Diving Fatalities 1969-2007

Uploaded by

Copyright:

Available Formats

American Cave Diving Fatalities 1969-2007

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

American Cave Diving Fatalities 1969-2007

Uploaded by

Copyright:

Available Formats

International Journal of Aquatic Research and Education

Volume 3 Number 2 Article 7

American Cave Diving Fatalities 1969-2007

Follow this and additional works at: https://scholarworks.bgsu.edu/ijare

International Journal of Aquatic Research and Education, 2009, 3, 162-177

American Cave Diving Fatalities 1969-2007

162by ScholarWorks@BGSU, 2009

Cave Diver Training in America

Cave Diving Fatality Reviews

Figure 1 — Cases reviewed and growth of the instrument.

166by ScholarWorks@BGSU, 2009

Figure 2 — Growth of classification categories by number of cases reviewed.

Published by ScholarWorks@BGSU, 2009 7

Table 1 Measures of Interrater Agreement

Figure 3 — Number of reported cave diving deaths by year.

Figure 4 — Proportion of cave diving deaths by trained divers, by year.

In addition to breaking the training rule, untrained divers were significantly

Published by ScholarWorks@BGSU, 2009 9

Table 2 Characteristics of Cave Divers by Training Status

Table 3 Cause of Death by Training Status

Table 4 Preceding Events by Training Status for Drowning

Table 5 Harmful Actions Preceding Air Depletion by Training Status

Table 6 Circumstances Preceding Getting Lost

Table 7 Circumstances That Preceded Diving With Insufficient Gas

172by ScholarWorks@BGSU, 2009

Published by ScholarWorks@BGSU, 2009 13

Published by ScholarWorks@BGSU, 2009 15

Walker, D. (2002). Report on Australian diving deaths 1994-1998. Ashburton, Victoria:

You might also like