Effect of experience and commercialisation on survival in Himalayan mountaineering: retrospective cohort study
Westhoff JL, Koepsell TD, Littell CT. Effect of experience and commercialisation on survival in Himalayan mountaineering: retrospective cohort study. BMJ 2012;344:e3782, doi:10.1136/bmj.e3782
Objectives—To determine whether previous Himalayan experience is associated with a decreased risk of climbing death, and whether mountaineers participating in commercial expeditions differ in their risk of death relative to those participating in traditional climbs.
Design—Retrospective cohort study.
Setting—Expeditions in the Nepalese Himalayan peaks, from 1 January 1970 to the spring climbing season in 2010.
Participants—23 995 non-porters venturing above base camp on 39 038 climbs, 23 295 on 8000 m peaks.
Results—After controlling for use of standard route, peak, age, season, sex, summit success, and year of expedition, increased Himalayan experience was not associated with a change in the odds of death (odds ratio 1.00, 95% confidence interval 0.96 to 1.05, P=0.904). Participation in a commercial climb was associated with a 37% lower odds of death relative to a traditional venture, although not significantly (0.63, 0.37 to 1.09, P=0.100). Choice of peak was clearly associated with altered odds of death (omnibus P<0.001); year of expedition was associated with a significant trend toward reduced odds of death (0.98, 0.96 to 0.99, P=0.011).
Conclusions—No net survival benefit is associated with increased Himalayan experience or participation in a traditional (versus commercial) venture. The incremental decrease in risk associated with calendar year suggests that cumulative, collective knowledge and general innovation are more important than individual experience in improving the odds of survival.
Why do the study?
Increasing numbers of climbers are attracted by the Himalayan mountains, and their climbing activities are associated with an extremely high risk of death.   It is important to provide accurate data on the risk of death from climbing, and any other risks. Such information is particularly relevant with the increasing number of alluring offers from commercial expedition organisers to participate in guided tours to the highest peaks in the world.
Several studies have tried to create a meaningful picture of mortality in the Himalayas.   However, those studies rarely evaluated the effects of previous Himalayan experience on the subsequent risk of death. Also, they did not compare the risk for traditional climbers with that of those participating in commercial expeditions. Thus Westhoff and colleagues focused on climbing experience and risk differences in traditional versus commercial ventures in their newly published study. The large sample and long observation period of 40 years (1970 to 2010) enabled them to analyse the effects of variables such as peak, route, altitude, season, and year.
What did the authors do?
The authors performed an observational study, or more specifically, a retrospective cohort study, which is an epidemiological study using data records that have previously been collected. To extrapolate the findings of such studies, it is important to draw a representative sample, or even better, to include each individual of the target population.
Westhoff and colleagues tried to do this by using a unique database created by Elizabeth Hawley. Hawley’s team has interviewed participants of almost every expedition to Nepal since 1963. In 2004, these archives were converted into the Himalayan Database.
In the study, observations were restricted to expeditions from 1970 to 2010 where climbers travelled above base camp, and excluded high altitude porters. Variables of main interest were previous Himalayan experience and participation in a commercial expedition. A total of 23 995 individuals and 39 038 climbs were included. Such a large sample strengthens study outcomes but causes problems with statistical analysis.
The authors assessed odds ratio of death using an analysis that allows for dependence. Such dependency is typically found in longitudinal data with repeated measurements—for instance, in this study, when analysing multiple observations in the same climber participating in more than one expedition. Thus the analysis required an adjustment for correlation. Standard logistic regression is a method for making predictions when there is a dichotomous dependent variable—that is, a dependent variable with two possible outcomes, in this case death or survival—and a set of explanatory variables, such as age, sex, season of climb, altitude, or peak. In this paper, the authors used an analysis which is based on the same model as standard logistic regression but also allows for dependence—that is, generalised estimating equation.
The Kaplan-Meier method was used to generate an estimate of cumulative mortality as a function of the number of expeditions. Such analysis allows the assessment of survival (or mortality) over time and is frequently used in medical research. It is often applied when patients drop out or when they are studied for different lengths of time (“censoring”). In this study, no censoring occurred.
What did the study find?
The study focused on the effect of previous Himalayan experience and participation in commercial expeditions. The analyses revealed that after controlling for several variables (use of standard route, peak, age, season of climb, sex, summit success, and year of expedition), Himalayan experience was not associated with a change in the risk of death (odds ratio 1.00, 95% confidence interval 0.96 to 1.05, P=0.90). Despite the observation that participation in a commercial climb was associated with a 37% lower risk of death compared with traditional climbing, this result was not statistically significant (odds ratio 0.63, 95% confidence interval 0.37 to 1.09, P=0.1).
Significantly different mortality rates were found for various Himalayan peaks. The unadjusted mortality (that is, with no compensation for factors such as age, season of climb, and so on) was lowest on Cho Oyu (0.5 % per climb) and highest on Annapurna (4 % per climb).
For a hypothetical Himalayan climbing career consisting of participation in 15 expeditions, a cumulative mortality of 20% was estimated (see figure). Importantly, the authors found a significant trend toward reduced risk of dying during the 40 year observation period (odds ratio 0.98, 95% confidence interval 0.96 to 0.99, P=0.01), meaning that the likelihood of death decreased from 3.0% per climb in the 1970s to 0.9% in the last decade. Other study findings concern the age and sex distributions of traditional and commercial expeditions, the frequencies of chosen routes and peaks, success rates (overall 31.4%), and causes of death. The overwhelming causes of death were falls (42.9%) and avalanches (28.9%). 1
What are the strengths and limitations of the study?
Retrospective studies are particularly susceptible to selection and information biases. Bias is defined as a deviation of findings from the truth. Whereas confounding bias can be controlled after completion of the study—for example, age, sex, season of climb, altitude—this is not possible for selection and information biases. Although most subjects at risk (climbers who were not porters and were climbing above base camp between 1970 and 2010) were recorded, one cannot entirely exclude a selection bias.
A potential information bias, however, might be more important since we do not have much information on the type of interviews, experience of interviewers, and related changes during the 40 year period, especially with regard to the definition of “commercial expeditions,” and also because of uncertainties with regard to the data converting process. Another weakness is that we do not know whether the many falls were primary causes of death or whether pre-existing high altitude illnesses were primary causes.
The large sample capturing nearly the entire at-risk population, the long observation period, and the sophisticated statistical analyses are undoubtedly strengths of the study.
What does this study mean?
This study confirmed the findings of Huey and colleagues that previous Himalayan experience does not alter the risk of death during a subsequent climb in the Himalayas, and that falls and avalanches were the most common causes of death. This is not unexpected. The risk of death from avalanche is difficult to estimate and remains high even for experienced mountaineers, and falls are associated with insufficient individual fitness and mountaineering skills. The acquisition of avalanche knowledge and fitness and mountaineering skills requires planned and long lasting—months to years—preparatory training. Thus, one would not expect a prior expedition alone to reduce the risk of death related to falls and avalanches.
This study found similar mortality for traditional climbers and climbers on commercial expeditions. However, although not statistically significant, participation in a commercial climb was associated with a 37% lower risk of death. Furthermore, a marked decrease in mortality over time was seen, and this decrease occurred when the number of participants in commercial expeditions increased dramatically. Thus, it would have been interesting to show results of cumulative mortality for a hypothetical Himalayan climbing career not only for the overall sample but for traditional climbers and participants of commercial expeditions separately as well.
It is important to highlight the extremely high risk of death on Himalayan peaks compared with other mountaineering activities. For example, the risk of dying when climbing Himalayan peaks ≥ 8000 m (544 deaths/1 million exposure days) is increased on average by a factor of 495 compared with downhill skiing (1.1 deaths/1 million exposure days) and by a factor of 56 compared with rock and ice climbing in the Alps (9.7 deaths/1 million exposure days). However, we must not forget that a large part of mortality is due to limited availability of rescue and medical services in remote areas. When evaluating whether to attempt high Himalayan peaks, mountaineers should be given access to quantitative data regarding their chance of reaching the summit and subsequent risk of death.
1University of Innsbruck, Austria
Correspondence to: Martin.Burtscher@uibk.ac.at
Competing interests: None declared.
Provenance and peer review: Commissioned; not externally peer reviewed.
- Firth PG, Zheng H, Windsor JS, Sutherland AI, Imray CH, Moore GW, et al. Mortality on Mount Everest, 1921-2006: descriptive study. BMJ 2008;337:a2654.
- Huey RB, Salisbury R, Wang JL, Mao M. Effects of age and gender on success and death of mountaineers on Mount Everest. Biol Lett 2007;3:498-500.
- Windsor JS, Firth PG, Grocott MP, Rodway GW, Montgomery HE. Mountain mortality: a review of deaths that occur during recreational activities in the mountains. Postgrad Med J 2009;85:316-21.
- Westhoff JL, Koepsell TD, Littell CT. The effects of experience and commercialisation on survival in Himalayan mountaineering: retrospective cohort study. BMJ 2012;344:e3782.
- Salisbury R. The Himalayan database: the expedition archives of Elizabeth Hawley. American Alpine Club, 2004.
- Zorn CJW. Generalized estimating equation models for correlated data: A review with applications. AJPS 2001;45:470-90.
- Burtscher M, Nachbauer W. Effects of training on the risk of avalanche fatality. In: Johnson RJ, ed. Skiing trauma and safety, 12th vol. American Society for Testing and Materials, 1999: 45-9.
- Burtscher M. Climbing the Himalayas more safely. BMJ 2012;344:e3778.
Cite this as: Student BMJ 2012;20:e4846