Development and Validation of a Dynamically Updated Prediction Model for Attrition From Marine Recruit Training

Dijksma, Iris; Hof, Michel H.; Lucas, Cees; Stuiver, Martijn M.

doi:10.1519/jsc.0000000000003910

Cited by 3 publications

(2 citation statements)

References 19 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Ideally, a study would report predictive performance for multiple models to avoid overfitting and dependence on one model, and use repeated measures to allow for prediction of dropout in advance. For a recent example in the context of the marine corps, see Dijksma et al (2022).…”

Section: Early Identification Of Dropouts During the Special Forces S...mentioning

confidence: 99%

Early Identification of Dropouts During the Special Forces Selection Program

Huijzer,

Blaauw,

de Wit

et al. 2024

Preprint

View full text Add to dashboard Cite

Special forces selection is a highly demanding process that involves exposure to high levels of psychological and physical stress resulting in dropout rates of up to 80%. To identify who likely drops out, we assessed a group of 249 recruits, every week of the program, on their experienced psychological and physical stress, recovery, self-efficacy, and motivation. Using both ordinary least squares regression and state-of-the-art machine learning models, we aimed to find the model that could predict dropout best. Furthermore, we inspected the best model to identify the most important predictors of dropout and to evaluate the predictive performance in practice. Via cross-validation, we found that linear regression performed best while remaining interpretable, with an Area Under the Curve (AUC) of 0.69. We also found that low levels of self-efficacy and motivation were significantly associated with dropout. Additionally, we found that dropout could often be predicted multiple weeks in advance and that the AUC score may underestimate the real-world predictive performance. Taken together, these findings offer novel insights in the use of prediction models on repeated measurements of psychological and physical processes, specifically in the context of special forces selection. This offers opportunities for early intervention and support, which could ultimately improve selection success rates.

show abstract

Section: Early Identification Of Dropouts During the Special Forces S...mentioning

confidence: 99%

Early Identification of Dropouts During the Special Forces Selection Program

Huijzer,

Blaauw,

de Wit

et al. 2024

Preprint

View full text Add to dashboard Cite

show abstract

“…Previously, we developed a time-updated prediction model including baseline characteristics, such as physical fitness and body weight, and weekly measured physical and mental health states, which will be reported elsewhere. In this model, of all statements, the motivation statement (“I felt motivated last week”) had the largest effect on dropout (Dijksma et al, 2021 ). The most frequently given reason for dropout was that the course was mentally too arduous.…”

Section: Introductionmentioning

confidence: 96%

Grit was not associated to dropout in Dutch Marine recruits

Dijksma¹,

Lucas²,

Stuiver³

2022

Military Psychology

Self Cite

View full text Add to dashboard Cite

Approximately half of all recruits drop out of Marine recruit training. Identifying associated and predisposing factors for dropout would be helpful to understand dropout patterns and induce preventive strategies. Grit has been suggested to be a predictor of who is likely to succeed and who is not. We aimed to investigate the association between baseline grit scores and dropout of Marine recruit training in the Netherlands Armed Forces. We performed an exploratory study using data of three platoons Marine recruit training of the Royal Netherlands Marine Corps. Individual grit levels were measured using the NL-Grit scale, including two subscales. The primary outcome of this study was successful completion or dropout of Marine recruit training. Data were available from 270 recruits, of whom 119 (44%) dropped out of training. The odds ratio for dropout were 1.01 (95% CI 0.84–1.21, p = .917) and 1.07 (95% CI 0.89–1.29, p = .481) per standard deviation increase of consistency of interests and perseverance of effort , respectively. Our study did not confirm the proposed association between baseline grit levels and dropout of Marine recruit training in Dutch Marine recruits using the NL-Grit scale.

show abstract

Predictors of combat training attrition in Israel Defense Forces soldiers

Gendler

Talmy

Shapiro

et al. 2023

Occupational Medicine

View full text Add to dashboard Cite

Background Attrition from combat service carries significant organizational and personal ramifications, but predicting factors associated with attrition remains challenging. Aims To evaluate medical and psychosocial factors associated with attrition from basic combat training (BCT) in the Israel Defense Forces (IDF). In addition, we identify subsets of the recruit population which exhibit certain trends in terms of medical corresponding with a high risk of attrition. Methods A cross-sectional study of IDF combat trainees undergoing infantry BCT between 2012 and 2017. Data were collected from the soldiers’ electronic medical and administrative records. We used multivariable logistic regression and the SAS® decision-tree tool to analyse key predictive factors for attrition. Results A total of 46 472 soldiers enlisted to BCT during the research period. The mean body mass index (BMI) was 21.8 (SD 3.54). The overall attrition rate was 10%. The following factors were associated with attrition from BCT: ethnicity (P < 0.01), BMI (P < 0.01), pre-enlisting motivation score (P < 0.01) and the number of mental health officer visits (P < 0.01). Using a decision-tree model, we found a high attrition rate among soldiers who had >5.2 to primary care physician visits (11% attrition rate versus 3%) or more than 11 sick leave days (59% versus 19%). Conclusions This study sheds light on unique measures relating to attrition. Attrition is associated with several demographic and psychosocial factors. Early prediction of motivation and monitoring of healthcare utilization may enable early identification and focused interventions targeting soldiers at high risk for attrition. These findings need to be further translated into actionable directives and further investigations.

show abstract

Development and Validation of a Dynamically Updated Prediction Model for Attrition From Marine Recruit Training

Cited by 3 publications

References 19 publications

Early Identification of Dropouts During the Special Forces Selection Program

Early Identification of Dropouts During the Special Forces Selection Program

Grit was not associated to dropout in Dutch Marine recruits

Predictors of combat training attrition in Israel Defense Forces soldiers

Contact Info

Product

Resources

About