Mental health after a mountain sport accident
Herein, we investigated patterns of mental health in a cross-sectional cohort of victims of mountain sport accidents treated in a tertiary trauma center in Tyrol, Austria. The analysis techniques encompassed semi-supervised PAM clustering, canonical hypothesis testing and machine learning with random forest, conditional random forest, neural network, recursive partitioning, support vector machines and elastic net regression algorithms.
By clustering in respect to measures of anxiety, depression, somatization, panic, resilience, sense of coherence, quality of life, post-traumatic growth and post-traumatic stress disorder signs, three subsets of accident victims were identified:
-
neutral cluster: individuals without symptoms of deteriorating mental health or post-traumatic symptom disorder but also without signs of post-traumatic growth
-
PTG cluster: named after post-traumatic growth which ist the key hallmark of the cluster.
-
PTS cluster: characterized by the highest intensity of post-traumatic stress disorder (PTSD) signs along with substantial anxiety, depression, somatic symptoms, loss of quality of life, reduced coherence and limited resilient coping
You may follow the analysis progress here.
The analysis pipeline requires some development packages, the easiest way to install them is to use devtools:
devtools::install_github('PiotrTymoszuk/ExDA')
devtools::install_github('PiotrTymoszuk/trafo')
devtools::install_github('PiotrTymoszuk/clustTools')
devtools::install_github('PiotrTymoszuk/soucer')
devtools::install_github('PiotrTymoszuk/figur')
devtools::install_github('PiotrTymoszuk/caretExtra')
devtools::install_github('PiotrTymoszuk/bootStat')
To launch the entire pipeline, source the exec.R file:
source('exec.R')
The pipeline results will be included in a future publication. To reference and use analysis results, please cite our GitHub repository; the data sources listed below and, if available, the publication. In any questions, please contact Dr. Katharina Hüfner.
The maintainer of the repository is Piotr Tymoszuk. Any data requests should be addressed to the senior author Dr. Katharina Hüfner.