Development and validation of a prediction model for the depressive symptom risk in commercial airline pilots.

Development and validation of a prediction model for the depressive symptom risk in commercial airline pilots - PMC Skip to main content An official website of the United States government Here's how you know Here's how you know Search Log in Primary site navigation Search Log in Search PMC Full-Text Archive Search in PMC Journal List User Guide As a library, NLM provides access to scientific literature. Inclusion in an NLM database does not imply endorsement of, or agreement with, the contents by NLM or the National Institutes of Health. Learn more: PMC Disclaimer | PMC Copyright Notice EPMA J . 2025 Apr 3;16(2):285–298. doi: 10.1007/s13167-025-00408-5 Development and validation of a prediction model for the depressive symptom risk in commercial airline pilots Jie Zhang 1, 2, 3 , Xuhua Chen 3 , Lin Zhang 4 , Haodong Qi 5 , Erliang Zhang 3 , Minzhi Chen 3 , Yiran Wang 6 , Yunfei Li 7 , Yan Chen 4 , Qingqing Duan 4 , Feng Shi 4 , Linlin Wang 4 , Qingqing Jin 4 , Bin Ren 4 , Yong Lu 8 , Ya Su 9 , Mi Xiang 1, 2, 3, ✉ Author information Article notes Copyright and License information 1 International Peace Maternity and Child Health Hospital, School of Medicine, Shanghai Jiao Tong University, Shanghai, China 2 Shanghai Key Laboratory of Embryo Original Diseases, Shanghai, China 3 School of Public Health, Shanghai Jiao Tong University School of Medicine, No. 227, South Chongqing Road, Shanghai, China 4 CAAC East China Aviation Personnel Medical Appraisal Center, Shanghai, China 5 Malmö Institute for Studies of Migration, Diversity and Welfare, Malmö University, Malmö, Sweden 6 College of Computing, City University of Hong Kong, Hong Kong, China 7 Department of Neurobiology, Care Sciences and Society, Karolinska Institute, Stockholm, Sweden 8 Radiology Department, Clinical Neuroscience Center, Rui Jin Hospital, School of Medicine, Shanghai Jiao Tong University, Shanghai, China 9 School of Nursing, Shanghai Jiao Tong University School of Medicine, No. 227, South Chongqing Road, Shanghai, China ✉ Corresponding author. Received 2024 Oct 18; Accepted 2025 Mar 26; Collection date 2025 Jun. © The Author(s), under exclusive licence to European Association for Predictive, Preventive and Personalised Medicine (EPMA) 2025. Springer Nature or its licensor (e.g. a society or other partner) holds exclusive rights to this article under a publishing agreement with the author(s) or other rightsholder(s); author self-archiving of the accepted manuscript version of this article is solely governed by the terms of such publishing agreement and applicable law. PMC Copyright notice PMCID: PMC12106183  PMID: 40438496 Abstract Background/aims Shift workers, such as medical personnel, and pilots, are facing an increased risk of depressive symptoms. Depressive symptoms significantly impact an individual’s quality of life and affect work performance, decision-making abilities, and overall public safety. This study aims to establish a multidimensional depressive symptom prediction model based on a large sample of commercial airline pilots to facilitate early identification, prevention, and personalized intervention strategies. Methods This population-based study included 11,111 participants, with 7918 pilots in the training set and 3193 pilots in the external validation set. Depressive symptom severity was assessed using the Patient Health Questionnaire-9 (PHQ-9). Physiological, psychological, and lifestyle factors potentially associated with depressive symptom risk were collected. The optimal predictors for model development were selected using the Boruta algorithm combined with the LASSO method, and a nomogram was developed using multivariate logistic regression to predict depressive symptoms in pilots. The model performance was evaluated using Receiver Operating Characteristic (ROC) curves, calibration curves, and accuracy measures, such as the Brier score and Spiegelhalter z -test. Additionally, decision curve analysis (DCA) was performed to assess the model's clinical utility. Results A total of 7918 pilots were included in the training set and 3193 were included in the external validation set. Five characteristic indicators were selected based on their significance in the prediction of depressive symptom risk: living statu s , alcohol drinking, family history of mental health disorder, subjective health, and subjective sleep quality. The model showed acceptable overall discrimination (AUC train = 0.836, 95%CI 0.818 to 0.854; AUC validation = 0.840, 95%CI 0.811 to 0.868) and calibration (Brier score train = 0.048; Brier score validation = 0.051). The decision curve analysis showed that the net benefit was superior to intervening on all participants or not intervening on all participants. Conclusions This study provides a reliable tool for early prediction and customized management of depressive symptoms among commercial airline pilots. This approach promotes the development of the field by transitioning from passive mental health care to active mental health prevention, emphasizing personalized prevention strategies. Keywords: Commercial airline pilots, Shift workers, Mental health, Depressive symptom risks, Personalized intervention strategy, Prediction model, Predictive Preventive Personalized Medicine (3PM / PPPM) Introduction The global prevalence of depression and depressive symptoms has been increasing in recent decades and has become one of the most serious mental disorders [ 1 ]. The GBD 2019 showed that depression has become one of the most disabling mental disorders. Previous studies found that the age-standardized prevalence of depressive symptoms in mainland China was 15.81% in 2019 [ 2 ]. Depression is a disease or mood disorder with a variety of depressive symptoms. Depending on the number and severity of depressive symptoms, depression can be classified as mild, moderate, or severe [ 3 , 4 ]. Depression can affect emotional and psychological health, cognition, behavior, and physical health, and may even lead to suicide [ 5 , 6 ]. It is worth noting that as shift workers, the prevalence of depression or depressive symptoms in pilots is higher than that in the general public. A recent study showed that the prevalence of depressive symptoms in Saudi Arabian Airlines pilots was as high as 40.6% [ 7 ]. In recent years, research on pilot mental health has received increasing attention due to its profound impact on flight safety. Pilots’ mental health is of vital importance as it directly affects their cognitive function, decision-making ability, situational awareness, and overall performance during flight operations, thereby endangering flight safety [ 8 ]. Despite facing unique job-related risks, pilots remain an under-researched group. In particular, research on mental health disorders among commercial pilots is limited. Factors such as age, gender, flight time, job stress and subjective sleep quality, alcohol consumption, physical activity level, burnout, history of depression, tenure, education, marital status, and role can affect their mental health and lead to depressive symptoms [ 8 – 10 ]. Maintaining pilot mental health is critical not only to ensuring the well-being of these individuals but also to safeguarding the safety of passengers and crew members involved in air travel [ 11 ]. High-risk status for depressive symptoms is a type of Suboptimal Health Status (SHS), which marks an early stage in the onset of depression and provides an opportunity for primary prevention strategies. As a subclinical and reversible stage of chronic diseases, SHS significantly contributes to predicting and preventing chronic diseases within the framework of PPPM [ 12 ]. Primary prevention, such as risk screening risk prediction, and identification, and secondary prevention, such as routine clinical screening and early intervention, may be applied to mitigate the risk of adverse outcomes that may be caused by depression. However, depression is a multifactorial mental health problem which includes demographic, physiological, psychological, and lifestyle behaviors factors and the specific key variables that influence the incidence of depression are different in different populations [ 13 , 14 ]. To our knowledge, there is currently no risk prediction model for depressive symptoms in commercial airline pilots, which may hinder the early prevention and intervention of depressive symptoms in pilots. Working hypothesis Although pilots receive regular physical and mental health checks, preventive and predictive measures are still crucial to identify stressed individuals at risk of transitioning from health to illness. Pilots are often required to work night and weekend shifts, which makes them more susceptible to circadian rhythm disruption and altered sleep patterns [ 15 ]. Previous research has found that irregular working hours can lead to a disruption of one’s morningness-eveningness preferences or chronotype, resulting in circadian disruption, which has been proved to be related to depression [ 16 ]. In addition, pilots often work in challenging environments where they must make critical decisions quickly [ 17 ]. As a result, pilots may experience sleep deprivation, stress, and circadian rhythm disruption. These issues have been linked to a variety of health problems, including depressive symptoms [ 18 – 20 ]. From a preventive and personalized medicine (PPPM) perspective, it is necessary to develop a rapid, reliable, and cost-effective predictive model for commercial airline pilots. This model should be designed specifically for pilots and provide a new risk prediction solution using easily accessible predictors. As emphasized in the European Predictive, Preventive and Personalized Medicine Association (EPMA) white paper, identifying pilots at high risk for depressive symptoms is a key step in preventing the negative consequences of depressive symptoms. [ 21 ] Depressive symptoms include cognitive, affective, and physical symptoms, and the influencing factors vary among different populations. We hypothesize that multiple factors contribute to depressive symptoms in pilots, including social and psychological factors [ 22 , 23 ]. To enable more personalized preventive measures, assessing changes in different functional dimensions is also important in risk prediction. Nomograms are powerful tools for clinical and nonclinical studies because they can accurately predict the associations between multiple variables and outcome measures (e.g., incidence and recurrence rates) by converting traditional statistical prediction models into visual graphs [ 24 ]. The Patient Health Questionnaire-9 (PHQ-9) is a valuable tool for assessing depressive symptoms and is primarily used for rapid screening and assessment of current symptoms [ 25 ]. Although it effectively identifies the presence and severity of depressive symptoms, it does not predict the future risk of depressive symptoms. From the perspective of PPPM/3PM, more comprehensive and accurate predictions are essential. Therefore, the goal of this study was to develop a multidimensional risk prediction model for depressive symptoms in commercial airline pilots. This model could be used to guide early intervention for pilots and enable predictive, preventive, and personalized medicine. Methods Study design and study population Our study is based on the Civil Aviation Health Cohort of China (CAHCC), more details about the CAHCC study can be found in previous publications [ 26 ]. All aircrew members participating in the CAHCC were active employees who met the active duty requirement and were excluded from being on long-term leave for any reason (e.g., mental disorder diagnosis, pregnancy, and other unfit to fly status). We utilized the personnel contact list provided by the airlines and distributed an online survey questionnaire link to all registered phone numbers via the Wenjuanxing platform ( https://www.wjx.cn/ ). Participants accessed the survey by clicking on the provided link, which directed them to the survey’s landing page. Before completing the questionnaire, all participants provided informed consent, acknowledging their voluntary participation and understanding of the study’s purpose. As of now, CAHCC has completed two rounds of data collection. For the purpose of this study, only pilots were included to analysis. During the baseline data collection (December 2022 to March 2023), a total of 22,617 questionnaires were obtained, from which 7918 pilots met the inclusion criteria and were included in the training set. In the first follow-up (November to December 2023), a total of 29,202 questionnaires were collected, including 3193 newly enrolled pilots, who were included in the external validation set, as illustrated in Fig. 1 . The training set is used for risk factor selection and prediction model building, while the validation set is used to evaluate the performance. Fig. 1. Open in a new tab The flow chart of the whole experiment Ethics statement This study was approved by the Ethics Committee of Civil Aviation Shanghai Hospital (No. 2021–7). All ethics procedures and governance requirements were adhered to for each interviewed participant. Measurement The PHQ-9 consists of 9 items of common symptoms of depression over the past 2 weeks on a scale of 0 to 27. A cut-off score of 10 or more indicates detection of depressive symptoms or possible depression [ 27 , 28 ]. For this study, we collected data based on previous research identifying potential factors associated with the risk of depressive symptoms among pilots. The collected data included the following domains: demographic characteristics (gender, age, education level, marital status, living status, working years, and flight hours), physical health (subjective health, pain, and skin diseases), mental health (stressful events, family history of mental health disorder), and lifestyle behaviors (cigarette smoking, alcohol drinking, coffee drinking, sleep duration, subjective sleep quality, physical activity, and sedentary time). Sleep duration and subjective sleep quality were measured using two self-report items: “During the past month, how many hours of actual sleep did you get at night? (This may be different than the number of hours you spent in bed.)” and “During the past month, how would you rate your sleep quality overall?” [ 29 ]. Sleep duration was categorized as < 7 h per day and ≥ 7 h per day [ 30 ]. Physical activity was assessed using the International Physical Activity Questionnaire (IPAQ) and was categorized based on a weekly activity level of 150 min: < 150 min/week and ≥ 150 min/week [ 31 ]. Sedentary time was categorized based on a self-reported daily sitting duration of 8 h: < 8 h and ≥ 8 h [ 32 ]. Statistical analysis To identify demographic characteristics, descriptive statistics were reported as frequencies and proportions for categorical variables and medians and interquartile ranges (IQRs) for continuous variables. In the univariate analysis, the chi-square test or Fisher’s exact test was used to analyze the categorical variables, while the Kruskal–Wallis test was used to examine the continuous variables. Feature selection is a crucial step in building predictive models. Therefore, we employed the Boruta algorithm in conjunction with LASSO to perform variable selection. The Boruta algorithm identifies feature importance by comparing the Z -scores of each real feature with the Z -scores of corresponding “shadow features.” Shadow features are copies of the real features, created by randomly shuffling, thereby eliminating the correlation between shadow features and the outcome variable. In each iteration, Z -scores for each feature (including both real and shadow features) are obtained from the random forest model. After multiple iterations, if the Z -score of a real feature is greater than the maximum Z -score among shadow attributes, the feature is considered “important” [ 33 ]. We employed the Boruta algorithm with parameter settings of “ P value = 0.01” and “maxRuns = 500”. Subsequently, the variables identified as important by the Boruta algorithm were further refined using the least absolute shrinkage and selection operator (LASSO) to eliminate collinearity. Through tenfold cross-validation, we selected the min lambda corresponding to non-zero coefficient variables, which were then included in the final multivariate logistic regression model, and a nomogram was constructed. Nomogram validation consisted of discrimination, accuracy, and clinical validity. The performance of the nomogram was appraised using the receiver operating characteristic (ROC) curve and calibration curve, with the area under the ROC curve (AUC) ranging from 0.5 to 1 (i.e., no discriminant to complete discriminant). Calibration accuracy was assessed using visual inspection, Brier score and P value of the Spiegelhalter z -test [ 34 ]. In addition, a decision curve analysis (DCA) was performed to assess the clinical utility of the nomogram by quantifying the net benefits at different threshold probabilities [ 35 ]. Results with a p-value of < 0.05 were deemed significant, and all tests were 2-tailed. All statistical analyses were performed using the R software (version 4.2.3). Results Participant characteristics A total of 11,111 civil aviation pilots were included in our study, of which 7918 were included in the training set and 3,193 in the external validation set. In the training set, 7824 (98.8%) pilots were male, with a median age of 32 years (IQR: 29–38). Significant differences ( P < 0.05) were observed between pilots with and without depressive symptoms in the training set concerning living status, skin diseases, pain, subjective health, stressful events, family history of mental health disorder, cigarette smoking, alcohol drinking, sleep duration, sleep quality, physical activity, and sedentary time. Additionally, a comparison between the baseline data of the training and validation sets revealed no significant differences ( P > 0.05) except for gender, education level, skin diseases, pain, and subjective sleep quality, indicating effective external validation of our data. In our study population, the depressive symptom detection rate in the training set was 6.0% (475/7918), and a rate of 6.4% (203/3193) in the external validation set. We presented the basic characteristics of both the training set and external validation set in our study population in Table 1 . Table 1. Baseline characteristics of all pilots in the training set and external validation set Characteristics Training set External validation set P value b Overall PHQ-9 < 10 PHQ-9 ≥ 10 P value a Overall PHQ-9 < 10 PHQ-9 ≥ 10 P value a N = 7918 N = 7443 N = 475 N = 3193 N = 2990 N = 203 Gender (%) 0.071 0.588 0.867 Male 7824 (98.8) 7350 (98.8) 474 (99.8) 3157 (98.9) 2955 (98.8) 202 (99.5) Female 94 (1.2) 93 (1.2) 1 (0.2) 36 (1.1) 35 (1.2) 1 (0.5) Age (median [IQR]) 32 [29, 38] 32 [29, 38] 33 [29, 37] 0.598 31 [27, 37] 31 [27, 37] 31 [28, 36] 0.616 < 0.001 Education level (%) 0.925 0.023 0.864 Below Bachelor’s degree 180 (2.3) 170 (2.3) 10 (2.1) 75 (2.3) 65 (2.2) 10 (4.9) Bachelor’s degree or above 7738 (97.7) 7273 (97.7) 465 (97.9) 3118 (97.7) 2925 (97.8) 193 (95.1) Marital status (%) 0.172 0.755 0.001 Single 1533 (19.4) 1430 (19.2) 103 (21.7) 714 (22.4) 667 (22.3) 47 (23.2) Married or stable partner 6257 (79.0) 5896 (79.2) 361 (76.0) 2420 (75.8) 2269 (75.9) 151 (74.4) Divorced or widowed 128 (1.6) 117 (1.6) 11 (2.3) 59 (1.8) 54 (1.8) 5 (2.5) Living status (%) < 0.001 < 0.001 0.001 Live alone 1393 (17.6) 1284 (17.3) 109 (22.9) 480 (15.0) 434 (14.5) 46 (22.7) Live with strangers 303 (3.8) 252 (3.4) 51 (10.7) 120 (3.8) 92 (3.1) 28 (13.8) Live with friends or colleagues 957 (12.1) 907 (12.2) 50 (10.5) 451 (14.1) 426 (14.2) 25 (12.3) Live with family 5265 (66.5) 5000 (67.2) 265 (55.8) 2142 (67.1) 2038 (68.2) 104 (51.2) Working years (median [IQR]) 7 [4, 12] 7 [4, 12] 8 [4, 12] 0.149 6 [3, 13] 6 [3, 13] 6 [3, 10] 0.413 < 0.001 Flight hours (median [IQR]) 4000 [1200, 9000] 4000 [1200, 9000] 4800 [1500, 9000] 0.224 4000 [800, 9900] 4000 [781, 10,000] 3500 [1200, 7690] 0.972 0.020 Skin diseases (%) < 0.001 < 0.001 0.116 No 7047 (89.0) 6676 (89.7) 371 (78.1) 2875 (90.0) 2719 (90.9) 156 (76.8) Yes 871 (11.0) 767 (10.3) 104 (21.9) 318 (10.0) 271 (9.1) 47 (23.2) Pain (%) < 0.001 < 0.001 0.248 No 3175 (40.1) 3116 (41.9) 59 (12.4) 1319 (41.3) 1300 (43.5) 19 (9.4) Yes 4743 (59.9) 4327 (58.1) 416 (87.6) 1874 (58.7) 1690 (56.5) 184 (90.6) Subjective health (%) < 0.001 < 0.001 < 0.001 Poor 624 (7.9) 460 (6.2) 164 (34.5) 243 (7.6) 177 (5.9) 66 (32.5) Moderate 3157 (39.9) 2908 (39.1) 249 (52.4) 1140 (35.7) 1030 (34.4) 110 (54.2) Good 4137 (52.2) 4075 (54.7) 62 (13.1) 1810 (56.7) 1783 (59.6) 27 (13.3) Stressful event (%) < 0.001 < 0.001 < 0.001 No 5797 (73.2) 5581 (75.0) 216 (45.5) 2519 (78.9) 2419 (80.9) 100 (49.3) Yes 2121 (26.8) 1862 (25.0) 259 (54.5) 674 (21.1) 571 (19.1) 103 (50.7) Family history of mental health disorder (%) < 0.001 < 0.001 0.001 No 7737 (97.7) 7313 (98.3) 424 (89.3) 3151 (98.7) 2965 (99.2) 186 (91.6) Yes 181 (2.3) 130 (1.7) 51 (10.7) 42 (1.3) 25 (0.8) 17 (8.4) Cigarette smoking (%) < 0.001 0.590 0.001 Never 4997 (63.1) 4737 (63.6) 260 (54.7) 2010 (63.0) 1887 (63.1) 123 (60.6) Have quit 847 (10.7) 789 (10.6) 58 (12.2) 275 (8.6) 259 (8.7) 16 (7.9) Yes 2074 (26.2) 1917 (25.8) 157 (33.1) 908 (28.4) 844 (28.2) 64 (31.5) Alcohol drinking (%) < 0.001 0.001 0.043 No 4350 (54.9) 4151 (55.8) 199 (41.9) 1686 (52.8) 1602 (53.6) 84 (41.4) Yes 3568 (45.1) 3292 (44.2) 276 (58.1) 1507 (47.2) 1388 (46.4) 119 (58.6) Coffee drinking (%) 0.147 0.925 < 0.001 No 2922 (36.9) 2762 (37.1) 160 (33.7) 1024 (32.1) 960 (32.1) 64 (31.5) Yes 4996 (63.1) 4681 (62.9) 315 (66.3) 2169 (67.9) 2030 (67.9) 139 (68.5) Sleep duration (%) < 0.001 < 0.001 < 0.001 < 7 h 1623 (20.5) 1423 (19.1) 200 (42.1) 524 (16.4) 433 (14.5) 91 (44.8) ≥ 7 h 6295 (79.5) 6020 (80.9) 275 (57.9) 2669 (83.6) 2557 (85.5) 112 (55.2) Subjective sleep quality (%) < 0.001 < 0.001 0.073 Bad 1250 (15.8) 983 (13.2) 267 (56.2) 460 (14.4) 359 (12.0) 101 (49.8) Good 6668 (84.2) 6460 (86.8) 208 (43.8) 2733 (85.6) 2631 (88.0) 102 (50.2) Physical activity (%) < 0.001 < 0.001 < 0.001 < 150 min/week 2445 (30.9) 2246 (30.2) 199 (41.9) 499 (15.6) 447 (14.9) 52 (25.6) ≥ 150 min/week 5473 (69.1) 5197 (69.8) 276 (58.1) 2694 (84.4) 2543 (85.1) 151 (74.4) Sedentary time (%) < 0.001 < 0.001 < 0.001 < 8 h 7041 (88.9) 6666 (89.6) 375 (78.9) 3020 (94.6) 2841 (95.0) 179 (88.2) ≥ 8 h 877 (11.1) 777 (10.4) 100 (21.1) 173 (5.4) 149 (5.0) 24 (11.8) Open in a new tab a Comparison of depression and non-depression pilot groups’ baseline characteristics in the training and external validation sets b Comparison of the training and external validation sets’ baseline characteristics Characteristic selection and predictors of depression The Boruta method, after 500 iterations, confirmed 10 important variables out of 19. Listed in ascending order based on Z -scores, these variables are coffee drinking, alcohol drinking, living status, marital status, family history of mental health disorder, subjective health, working years, age, flight hours, and subjective sleep quality as shown in Fig. 2 a, b. Then these 10 variables were subjected to LASSO method to simplify the model, resulting in the selection of 5 variables: living status, alcohol drinking, family history of mental health disorder, subjective health, and subjective sleep quality as shown in Fig. 2 c, d. Fig. 2. Open in a new tab Feature selection by the Boruta algorithm and LASSO method. a The process of feature selection. b The value evolution of Z -score in the screening process. c LASSO coefficient profiles of the features. d Optimal parameter (lambda) selection in the LASSO model used 10-folds cross-validation via minimum criteria. The horizontal axis shows the name of each variable and the number of iterations for the algorithm in Fig. 2a, b, respectively. While the vertical axis represents the Z -value of each variable. The green boxes and lines represent confirmed variables, the yellow ones represent tentative attributes, and the red ones represent rejected variables in the model calculation Nomogram development The final set of 5 selected variables were included in a multivariate logistic regression. The results indicated that living with strangers (OR = 2.526, 95%CI: 1.751–3.645), moderate subjective health (OR = 3.545, 95%CI: 2.641–4.759), poor subjective health (OR = 10.575, 95%CI: 7.571–14.772), family history of mental health disorder (OR = 4.464, 95%CI: 2.830–6.127), alcohol drinking (OR = 1.422, 95%CI: 1.160–1.743), and poor subjective sleep quality (OR = 4.255, 95%CI: 3.444–5.527) were identified as risk factors for depression among pilots as shown in Table 2 . We generated a visual nomogram based on the final logistic regression model to illustrate the predictive model as shown in Fig. 3 . Table 2. Results of multivariate logistic regression for training set Characteristics N Event N OR Lower 95%CI Upper 95%CI Living status Live with family 5265 265 Ref – – Live alone 1393 109 1.138 0.885 1.463 Live with strangers 303 51 2.526 1.751 3.645 Live with friends or colleagues 957 50 1.181 0.846 1.648 Subjective health Good 4137 62 Ref – – Moderate 3157 249 3.545 2.641 4.759 Poor 624 164 10.575 7.571 14.772 Family history of mental health disorders No 7737 424 Ref – – Yes 181 51 4.464 2.830 6.127 Alcohol drinking No 2922 160 Ref – – Yes 4996 315 1.422 1.160 1.743 Subjective sleep quality Good 6668 267 Ref – – Bad 1250 208 4.255 3.444 5.527 Open in a new tab OR odds ratio, CI confidence interval Fig. 3. Open in a new tab Nomogram of depressive symptom risk prediction for pilots. Each variable is assigned a score in the “Points” axis. Locate the sum of all the single scores combined in the “Total Points” axis: the point identified by a line drawn downwards on the “Risk of depressive symptoms” axis corresponds to the estimated probability of depression risk anticipated by the nomogram Validation of nomogram Discrimination AUC values were calculated to evaluate the discrimination of the predictive model by examining the depressive symptom risk in pilots in both training and validation sets. The AUC of the logistic regression model of the training set was 0.836 (95%CI, 0.818–0.854) and that in the external validation set was 0.840 (95%CI, 0.811–0.868) as shown in Fig. 4 a, b. These data indicate that the nomogram can correctly identify pilots at higher risk of depressive symptoms with good discriminatory ability and predictive value. Fig. 4. Open in a new tab ROC curves, calibration curves, and decision curve analysis for the risk of depressive symptoms. a, c, e Training set. b, d, f Validation set. ROC, receiver operating characteristic curve; AUC, area under the ROC curve Calibration of the predictive model Upon visual inspection of the calibration plot, it was observed that the predicted and observed probabilities overlapped well without relevant overestimations or underestimations. Brier score was 0.048 and 0.051, and the Spiegelhalter z -test was not significant ( P = 0.732 and 0.734) in the training and external validation set. Based on the results, it can be concluded that the nomogram is valid for use in the validation set. The calibration curve of the prediction model closely resembles the ideal curve, suggesting that the predicted outcomes are consistent with the actual findings as shown in Fig. 4 c, d. Evaluation of clinical validity The clinical validity of the model was assessed using the DCA method, and the DCA curves related to the nomogram are shown in as shown in Fig. 4 e, f. Threshold ranges for DCA were derived from the training and external validation sets based on the sensitivity and specificity of the prediction model. Intervention was performed on patients whose assessed risk was within the threshold range. The net benefit is superior to intervening on all participants or not intervening on all participants. Discussion This study included a total of 11,111 pilots. Among them, a training set of 7918 pilots were used to develop a nomogram that predicts depressive symptom risk in commercial airline pilots based on physiological, psychological, and behavioral factors. Living status, subjective sleep quality, family history of mental health disorders, subjective health, and alcohol drinking were factored into the prediction model. The model was then independently and externally validated using data from an additional 3193 pilots, which shows acceptable overall discrimination and calibration. This study found that subjective health is a strong predictor of depressive symptoms. Lower subjective health increases the risk of depressive symptoms. One possible explanation for this finding is that pilots with lower subjective health might worry more about their physical condition and struggle to keep a positive mood, which could lead to depression over time. Previous research has proved that subjective health is related to depression among workers [ 36 ]. Moreover, job-related diseases of pilots such as tiredness, sleep problems, bloating, lower back pain, headaches, and neck pain might lower their quality of life and potentially lead to depression [ 37 ]. In addition, the challenging nature of this position may exacerbate the negative effects of lower subjective health for pilots compared [ 38 , 39 ]. Psychological problems can manifest as physical symptoms, and physical problems can affect mental health [ 40 ]. Thus, by assessing subjective health through regular screenings, we can identify at-risk pilots early, enabling timely interventions. Family history of mental health disorder is also one of the predictive factors for depressive symptoms. Research has shown that depression is a disease with familial aggregation [ 41 , 42 ]. Moreover, subjective sleep quality is important in predicting depressive disorder among pilots. Since working places may be far from pilots’ homes, night shift may disturb the individual due to the long commute [ 43 ]. The irregularity of working schedules may also lead to sleep deprivation for pilots. As a result, some pilots suffer from low subjective sleep quality, increasing the risk of depressive symptoms [ 44 ]. Thus, managing low subjective sleep quality is essential to prevent depressive symptoms [ 45 ]. In our study, we found that the living status of pilots is a significant predictor of depressive symptoms. Some of the pilots need to live in a dormitory or rent with a stranger [ 46 , 47 ]. Sharing a room with a stranger is the most prominent factor, much higher than living with friends or living alone. On the other hand, living with family showed the lowest risk of depressive symptoms. The social support that pilots receive from their roommates has a significant impact on their mental health, which correlates with previous studies [ 48 , 49 ]. Furthermore, we discovered a link between alcohol consumption and depressive symptoms. This finding is consistent with previous research on various populations across different occupations and age groups [ 50 , 51 ]. Pilots may drink alcohol due to their backgrounds, work pressures, and family caregiving responsibilities [ 52 , 53 ]. The cumulative effects of drinking can also lead to physical and mental health consequences [ 54 , 55 ]. Our study highlights the urgent need to pay more attention to their drinking behaviors and give them appropriate health care. Our findings highlight the importance of subjective health status, family history of mental health disorders, and subjective sleep quality in predicting depressive symptom risk among pilots. In terms of prevention, our findings suggest that more health education and psychological support are needed for pilots to improve their subjective health status and prevent the occurrence of depression. This includes providing health promotion activities, stress management training, and psychological counseling services, as well as improving the working environment and conditions to reduce occupational-related diseases such as fatigue and sleep problems, which are important factors affecting subjective health and quality of life. For pilots with a family history of mental health disorders, targeted prevention measures such as regular mental health screening, stress management, and psychoeducation can be implemented. We advocate tailored interventions for the different health and psychological needs of pilots. For pilots who are under high stress due to occupational requirements, targeted workplace health promotion programs can be developed and implemented to reduce their psychological stress and improve their subjective health. Personalized psychological intervention plans can be designed based on the individual's family history, work environment, and personal preferences. Moreover, recent research has increasingly focused on biological markers, such as cortisol, as potential predictive tools for psychological health and performance in high-stress occupations such as commercial airline pilots [ 56 , 57 ]. Though our findings primarily focus on subjective health status, family history and subjective sleep quality as risk factors for depressive symptoms among pilots, biological discoveries could significantly complete these statistical findings. Future research should investigate how these psychosocial risk factors interact with biological markers to better predict and prevent depression in this high-risk population. Strengths Firstly, this study is the first large-scale, population-based survey conducted specifically on commercial airline pilots, aimed at developing a nomogram model to predict the risk of depressive symptoms. The large sample size provides strong representativeness and statistical power to the study. Secondly, our variable screening method, which combines the LASSO and Boruta algorithms, ensures not only the importance of the variables but also the simplicity of the model. Then, we comprehensively screened variables across four categories, ensuring that the model covers a wide range of factors influencing the risk of depressive symptoms. Furthermore, the data collected in the study were easy to gather, which not only improves the feasibility of data collection but also makes the model more practical and applicable in real-world applications. This convenience offers the possibility of further external validation of the model with pilots from other countries or regions in the future. Lastly, the study used an independent external set for rigorous validation, further enhancing the credibility of our findings. In the context of PPPM/3PM, our model fills a gap by providing pilots with a simple and accurate predictive tool. This model not only is helpful in predicting the risk of depressive symptoms but also provides essential information for personal mental health interventions, holding significant clinical and public health importance. Limitations However, this study does have some limitations. First, although the assessment in this study included multiple aspects of depressive symptoms, changes in effect or cognitive function may have been under-assessed. Therefore, incorporating these additional assessment indicators in future studies is essential to improve the accuracy and comprehensiveness of the prediction model. Moreover, our prediction model was based on self-report questions, so we cannot exclude the possibility of recall bias or response bias. Self-reported subjective sleep quality and sleep duration were assessed based on the past 4 weeks, but the PHQ-9 evaluated depressive symptoms over the past 2 weeks. Their assessment time windows are not fully aligned but are partially overlapping. Future studies could incorporate longitudinal sleep assessments to better capture changes in sleep patterns and their relationship with depressive symptoms over time. Second, in our study, the gender distribution of the pilot sample reflects the current reality in the aviation industry, where the number of male pilots is significantly greater than that of female pilots. However, this imbalance in the gender ratio may affect the accuracy of the data for female pilots. Third, as our study is based on cross-sectional observations, it lacks prospective insights into stress trajectories over time. Therefore, we cannot determine causal relationships between coping mechanisms and depressive symptoms in pilots. Future research could adopt longitudinal study designs to better understand the trajectories of mental health changes in pilots facing long-term work-related stress, as well as the role of coping mechanisms in this context. Last, depressive symptoms are complex and multifactorial that everyone is different. It may not be sufficient to use only a few variables to capture the heterogeneity and dynamics of depressive symptoms. In the future study, we will consider the inclusion of phenotyping, creation of individualize pilot profiles, pilot stratification linked to proposed risk mitigation measures and treatments tailored to the person these basic tools of PPPM. These tools can help to achieve more precise and individualized prevention and intervention strategies and can offer further hope for the future of primary prevention of depressive symptoms among pilots, thus leading to better PPPM. This will also help achieve the broader goal of transforming healthcare from a reactive model to PPPM, ultimately improving the quality of life of pilots in all industries. Conclusions and expert recommendations In this study, we developed a machine learning-based multidimensional risk prediction model to assess depressive symptom risk among pilots. Using a large set of 11,111 participants, we identified key factors of physiological, psychological, and lifestyle behavioral associated with depressive symptoms in this population. Through independent validation set, we highlight this model’s reliability and potential for widespread application. This nomogram improved the early identification of depressive symptom risk among pilots and laid the groundwork for developing targeted support strategies to enhance their mental health and well-being, thus facilitating preventive measures from a PPPM perspective. Our study lays the foundation for future studies to refine predictive models and explore additional factors influencing depressive symptom risk across different occupation groups. For the further application this prediction model in context of PPPM for depressive symptom risk management in commercial airline pilots, we recommend the following. Added value and paradigm shift This study marks a significant advancement in the field of occupational mental health by introducing a multidimensional risk prediction model based on machine learning. Our approach, which leveraged a comprehensive dataset of 11,111 participants, not only identified key physiological, psychological, and lifestyle behavioral factors associated with depressive symptoms in pilots, but also demonstrated the reliability of the model through independent validation. The model stands out for its potential for widespread application, enhancing early identification of depressive symptom risk and paving the way for tailored support strategies aimed at improving the mental health and well-being of pilots. By moving from a reactive to a proactive mental health stance, our study embodies a paradigm shift toward prediction, prevention, and personalized medicine (PPPM/3PM). It enables pilots to self-assess their depressive symptom risk based on a variety of factors, including living statu s , alcohol drinking, family history of mental health disorders, subjective health and subjective sleep quality, enabling timely interventions such as psychotherapy, lifestyle changes, stress management, or medication to improve mental health. Future directions Building on the foundation laid by this study, future research should focus on refining the prediction model and expanding its scope to include other factors that may affect depressive symptom risk in different occupational groups. This will involve continued monitoring and incorporation of new data to improve the accuracy and applicability of the model. For example, genetic underpinnings and possible biomarkers can further enhance our understanding of the pathogenesis of depressive symptoms and provide a scientific basis for the development of predictive tools. In addition, future monitoring can also be achieved in conjunction with wearable devices, sleep diaries, or mobile health applications to provide data support for early intervention. In addition, developing personalized intervention plans based on individual risk profiles is another area worth exploring. This may involve creating customized plans that combine psychotherapy, lifestyle changes, stress management techniques, and pharmacological interventions tailored to the specific needs of each pilot. This will help achieve the broader goal of transforming healthcare from a reactive model to a proactive, personalized, and preventive model, ultimately improving the quality of life of workers in all industries. Author contribution Drs Xiang, Su, and Lu had full access to all of the data in the study and took responsibility for the integrity of the data and the accuracy of the data analysis. Concept and design: J.Z., X.C., L.Z., Y.S., M.X.. Acquisition, analysis, or interpretation of data: All authors. Drafting of the manuscript: J.Z., X.C., Y.S.. Critical review of the manuscript for important intellectual content: L.Z., E.Z., M.C., Y.L., X.M.. Statistical analysis: J.Z., L.Z., Y.W.. Obtained funding: M.X.. Administrative, technical, or material support: M.X., Y.S.. Supervision: M.X., Y.S., Y.L.. All authors read and approved the final manuscript. Funding This project is funded by (1) the National Natural Science Foundation of China (No.: 71804110), (2) Shanghai Science and Technology Development Funds (No.: 21QA1405300), (3) Science Foundation of Ministry of Education of China (No.: 22YJAZH116), (4) Civil aviation safety capacity building project (No.: 251), and (5) the National Social Science Fund of China (No.24CSH108). Data availability No datasets were generated or analysed during the current study. Declarations Ethics approval This study was approved by the Ethics Committee of Civil Aviation Shanghai Hospital (No. 2021–7). All ethics procedures and governance requirements were adhered to for each interviewed participant. Consent to participate Informed consent was obtained from all participants. Competing interests The authors declare no competing interests. Footnotes Jie Zhang, Xuhua Chen, and Lin Zhang are made equal contributions to this study and are the co-first authors. Publisher's Note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. References 1. COVID-19 Mental Disorders Collaborators. Global prevalence and burden of depressive and anxiety disorders in 204 countries and territories in 2020 due to the COVID-19 pandemic. Lancet. 2020;2021(398):1700–12 [ DOI ] [ PMC free article ] [ PubMed ] 2. Liu J, Wu J, Wang J, Chen S, Yin X, Gong Y. Prevalence and associated factors for depressive symptoms among the general population from 31 provinces in China: the utility of social determinants of health theory. J Affect Disord. 2024;347:269–77. [ DOI ] [ PubMed ] [ Google Scholar ] 3. Robertson R, Robertson A, Jepson R, Maxwell M. Walking for depression or depressive symptoms: a systematic review and meta-analysis. Ment Health Phys Act. 2012;5:66–75. [ Google Scholar ] 4. Ayuso-Mateos JL, Nuevo R, Verdes E, Naidoo N, Chatterji S. From depressive symptoms to depressive disorders: the relevance of thresholds. Br J Psychiatry. 2010;196:365–71. [ DOI ] [ PubMed ] [ Google Scholar ] 5. Klein DN. Assessment of depression in adults and youth. Assessment. 2024;31:110–25. [ DOI ] [ PubMed ] [ Google Scholar ] 6. Deng M-G, Liu F, Liang Y, Wang K, Nie J-Q, Liu J. Association between frailty and depression: a bidirectional Mendelian randomization study. Sci Adv. 2023;9:eadi3902. [ DOI ] [ PMC free article ] [ PubMed ] 7. AlMuammar S, Alkhaldi R, Alsharif R, Allbdi D, Alasmari M, Alasmari B, et al. Navigating the skies: a cross-sectional study of depression among Saudi Arabian airline pilots. J Occup Med Toxicol. 2024;19:36. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 8. Newcomb RD, Steffen MW, Breeher LE, Sturchio GM, Murad MH, Wang Z, et al. Screening for depression in the occupational health setting. Occupational Medicine (Oxford, England). 2016;66:390. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 9. Wu AC, Donnelly-McLay D, Weisskopf MG, McNeely E, Betancourt TS, Allen JG. Airplane pilot mental health and suicidal thoughts: a cross-sectional descriptive study via anonymous web-based survey. Environ Health. 2016;15:121. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 10. Schonfeld IS, Bianchi R. From burnout to occupational depression: recent developments in research on job-related distress and occupational health. Front Public Health. 2021;9: 796401. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 11. Karkala A, Moschonas S, Sykas G, Karagianni M, Gilou S, Papaefthymiou O, et al. Sleep quality and mental health consequences of COVID-19 pandemic in the aviation community in Greece. J Occup Environ Med. 2022;64:e567–74. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 12. Wang W, Yan Y, Guo Z, Hou H, Garcia M, Tan X, et al. All around suboptimal health — a joint position paper of the Suboptimal Health Study Consortium and European Association for Predictive, Preventive and Personalised Medicine. EPMA J. 2021;12:403. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 13. Shen L, Xu X, Yue S, Yin S. A predictive model for depression in Chinese middle-aged and elderly people with physical disabilities. BMC Psychiatry. 2024;24:305. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 14. Kuang Y, Liao X, Jiang Z, Gu Y, Liu B, Tan C, et al. Federated learning-based prediction of depression among adolescents across multiple districts in China. J Affect Disord. 2025;369:625–32. [ DOI ] [ PubMed ] [ Google Scholar ] 15. Jeong Y-W, Kim J-H. Depression, anxiety, stress and anger of the cabin crew during the COVID-19 pandemic in South Korea. Healthcare (Basel). 2022;10:1952. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 16. Taillard J, Sagaspe P, Philip P, Bioulac S. Sleep timing, chronotype and social jetlag: Impact on cognitive abilities and psychiatric disorders. Biochem Pharmacol. 2021;191: 114438. [ DOI ] [ PubMed ] [ Google Scholar ] 17. He X, Nie X, Zhou R, Yang J, Wu R. The risk-taking behavioural intentions of pilots in adverse weather conditions: an application of the theory of planned behaviour. Ergonomics. 2023;66:1043–56. [ DOI ] [ PubMed ] [ Google Scholar ] 18. Riemann D, Krone LB, Wulff K, Nissen C. Sleep, insomnia, and depression. Neuropsychopharmacology. 2020;45:74–89. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 19. Park C, Rosenblat JD, Brietzke E, Pan Z, Lee Y, Cao B, et al. Stress, epigenetics and depression: a systematic review. Neurosci Biobehav Rev. 2019;102:139–52. [ DOI ] [ PubMed ] [ Google Scholar ] 20. Crouse JJ, Carpenter JS, Song YJC, Hockey SJ, Naismith SL, Grunstein RR, et al. Circadian rhythm sleep-wake disturbances and depression in young people: implications for prevention and early intervention. Lancet Psychiatry. 2021;8:813–23. [ DOI ] [ PubMed ] [ Google Scholar ] 21. Golubnitschaja O, Costigliola V. General Report & Recommendations in Predictive, Preventive and Personalised Medicine 2012: White Paper of the European Association for Predictive, Preventive and Personalised Medicine. EPMA J. 2012;3:14. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 22. Curran E, Perra O, Rosato M, Ferry F, Leavey G. Complex childhood trauma, gender and depression: patterns and correlates of help-seeking and maladaptive coping. J Affect Disord. 2021;292:603–13. [ DOI ] [ PubMed ] [ Google Scholar ] 23. Korja R, Nolvi S, Scheinin NM, Tervahartiala K, Carter A, Karlsson H, et al. Trajectories of maternal depressive and anxiety symptoms and child’s socio-emotional outcome during early childhood. J Affect Disord. 2024;349:625–34. [ DOI ] [ PubMed ] [ Google Scholar ] 24. Lv J, Ren H, Guo X, Meng C, Fei J, Mei H, et al. Nomogram predicting bullying victimization in adolescents. J Affect Disord. 2022;303:264–72. [ DOI ] [ PubMed ] [ Google Scholar ] 25. Costantini L, Pasquarella C, Odone A, Colucci ME, Costanza A, Serafini G, et al. Screening for depression in primary care with Patient Health Questionnaire-9 (PHQ-9): A systematic review. J Affect Disord. 2021;279:473–83. [ DOI ] [ PubMed ] [ Google Scholar ] 26. Zhang E, Li H, Han H, Wang Y, Cui S, Zhang J, et al. Dietary Rhythmicity and Mental Health Among Airline Personnel. JAMA Netw Open. 2024;7: e2422266. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 27. Dyball D, Bennett AN, Schofield S, Cullinan P, Boos CJ, Bull AMJ, et al. Mental health outcomes of male UK military personnel deployed to Afghanistan and the role of combat injury: analysis of baseline data from the ADVANCE cohort study. Lancet Psychiatry. 2022;9:547–54. [ DOI ] [ PubMed ] [ Google Scholar ] 28. Kroenke K, Spitzer RL, Williams JB. The PHQ-9: validity of a brief depression severity measure. J Gen Intern Med. 2001;16:606–13. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 29. Buysse DJ, Reynolds CF, Monk TH, Berman SR, Kupfer DJ. The Pittsburgh Sleep Quality Index: a new instrument for psychiatric practice and research. Psychiatry Res. 1989;28:193–213. [ DOI ] [ PubMed ] [ Google Scholar ] 30. Chunnan L, Shaomei S, Wannian L. The association between sleep and depressive symptoms in US adults: data from the NHANES (2007–2014). Epidemiol Psychiatr Sci. 2022;31: e63. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 31. Pearce M, Garcia L, Abbas A, Strain T, Schuch FB, Golubic R, et al. Association between physical activity and risk of depression. JAMA Psychiat. 2022;79:550–9. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 32. Zhou Q, Guo C, Yang X, He N. Dose–response association of total sedentary behaviour and television watching with risk of depression in adults: a systematic review and meta-analysis. J Affect Disord. 2023;324:652–9. [ DOI ] [ PubMed ] [ Google Scholar ] 33. Yue S, Li S, Huang X, Liu J, Hou X, Zhao Y, et al. Machine learning for the prediction of acute kidney injury in patients with sepsis. J Transl Med. 2022;20:215. [ DOI ] [ PMC free article ] [ PubMed ] 34. Demirjian S, Bashour CA, Shaw A, Schold JD, Simon J, Anthony D, et al. Predictive accuracy of a perioperative laboratory test-based prediction model for moderate to severe acute kidney injury after cardiac surgery. JAMA. 2022;327:956–64. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 35. Yu Y, Tan Y, Xie C, Hu Q, Ouyang J, Chen Y, et al. Development and validation of a preoperative magnetic resonance imaging radiomics-based signature to predict axillary lymph node metastasis and disease-free survival in patients with early-stage breast cancer. JAMA Netw Open. 2020;3: e2028086. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 36. Nakamura-Taira N, Izawa S, Yamada KC. Stress underestimation and mental health literacy of depression in Japanese workers: a cross-sectional study. Psychiatry Res. 2018;262:221–8. [ DOI ] [ PubMed ] [ Google Scholar ] 37. Omholt ML, Tveito TH, Ihlebæk C. Subjective health complaints, work-related stress and self-efficacy in Norwegian aircrew. Occup Med (Lond). 2017;67:135–42. [ DOI ] [ PubMed ] [ Google Scholar ] 38. Sweet E, Nandi A, Adam E, McDade T. The high price of debt: household financial debt and its impact on mental and physical health. Soc Sci Med. 2013;91:94–100. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 39. Beiter R, Nash R, McCrady M, Rhoades D, Linscomb M, Clarahan M, et al. The prevalence and correlates of depression, anxiety, and stress in a sample of college students. J Affect Disord. 2015;173:90–6. [ DOI ] [ PubMed ] [ Google Scholar ] 40. Salmon M, Sibeoni J, Harf A, Moro MR, Ludot-Grégoire M. Systematic review on somatization in a transcultural context among teenagers and young adults: Focus on the nosography blur. Front Psychiatry. 2022;13: 897002. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 41. van Dijk MT, Murphy E, Posner JE, Talati A, Weissman MM. Association of multigenerational family history of depression with lifetime depressive and other psychiatric disorders in children. JAMA Psychiat. 2021;78:1–11. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 42. Weersing VR, Shamseddeen W, Garber J, Hollon SD, Clarke GN, Beardslee WR, et al. Prevention of depression in at-risk adolescents: predictors and moderators of acute effects. J Am Acad Child Adolesc Psychiatry. 2016;55:219–26. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 43. Tse SL, Wang F, Li Z. P-194 Sex difference of negative emotion and contributing factors among Chinese nightshift workers. Occup Environ Med. 2021;78:A84–A84. [ Google Scholar ] 44. Jungquist CR, Mund J, Aquilina AT, Klingman K, Pender J, Ochs-Balcom H, et al. Validation of the behavioral risk factor surveillance system sleep questions. J Clin Sleep Med. 2016;12:301–10. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 45. Plante DT. The evolving nexus of sleep and depression. AJP. 2021;178:896–902. [ DOI ] [ PubMed ] [ Google Scholar ] 46. Zhao Y, Wang Y, Guo W, Cheng L, Tong J, Ji R, et al. Studies on the relationship between occupational stress and mental health, performance, and job satisfaction of Chinese civil aviation pilots. Aerospace. 2023;10:896. [ Google Scholar ] 47. Buchvold HV, Pallesen S, Waage S, Bjorvatn B. Shift work schedule and night work load: Effects on body mass index – a four-year longitudinal study. Scand J Work Environ Health. 2018;44:251–7. [ DOI ] [ PubMed ] [ Google Scholar ] 48. Almomani FM, Bani-issa W. The incidence of depression among residents of assisted living: prevalence and related risk factors. Clin Interv Aging. 2017;12:1645–53. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 49. Xu Q, Wu Y, Ji M, Wang M, Pan C, Ma J, et al. Proactive coping and mental health among airline pilots during China’s regular prevention and control of COVID-19: The role of perceived stress and social support. Front Public Health. 2022;10: 890145. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 50. Wu LI-TZY. Schlenger WILLIAME, Galvin DEBORAHM. The relationship between employment and substance use among students aged 12 to 17. Journal of Adolescent Health. 2003;32:5–15. [ DOI ] [ PubMed ] [ Google Scholar ] 51. Liu Q, Leng P, Gu Y, Shang X, Zhou Y, Zhang H, et al. The dose–effect relationships of cigarette and alcohol consumption with depressive symptoms: a multiple-center, cross-sectional study in 5965 Chinese middle-aged and elderly men. BMC Psychiatry. 2022;22:657. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 52. Grenard DL, Valencia EJ, Brown JA, Winer RL, Littman AJ. Impact of caregiving during emerging adulthood on frequent mental distress, smoking, and drinking behaviors: United States, 2015–2017. Am J Public Health. 2020;110:1853–60. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 53. Tao R, Jiang F, Min K, Liu T, Liu Y, Xia L, et al. Alcohol use among mental health professionals in China: a nationwide cross-sectional survey. Alcohol Alcohol. 2021;56:351–9. [ DOI ] [ PubMed ] [ Google Scholar ] 54. Bolstad I, Alakokkare A, Bramness JG, Rognli EB, Levola J, Mustonen A, et al. The relationships between use of alcohol, tobacco and coffee in adolescence and mood disorders in adulthood. Acta Psychiatr Scand. 2022;146:594–603. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 55. Wu J, Dong W, Pan X-F, Feng L, Yuan J-M, Pan A, et al. Relation of cigarette smoking and alcohol drinking in midlife with risk of cognitive impairment in late life: the Singapore Chinese Health Study. Age Ageing. 2019;48:101–7. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 56. Thomas L, Schwaninger A, Heimgartner N, Hedinger P, Hofer F, Ehlert U, et al. Stress-induced cortisol secretion impairs detection performance in x-ray baggage screening for hidden weapons by screening novices. Psychophysiology. 2014;51:912–20. [ DOI ] [ PubMed ] [ Google Scholar ] 57. Karabatsiakis A, de Punder K, Salinas-Manrique J, Todt M, Dietrich DE. Hair cortisol level might be indicative for a 3PM approach towards suicide risk assessment in depression: comparative analysis of mentally stable and depressed individuals versus individuals after completing suicide. EPMA J. 2022;13:383–95. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] Associated Data This section collects any data citations, data availability statements, or supplementary materials included in this article. Data Availability Statement No datasets were generated or analysed during the current study. Articles from The EPMA Journal are provided here courtesy of Springer ACTIONS View on publisher site PDF (1.2 MB) Cite Collections Permalink RESOURCES Similar articles Cited by other articles Links to NCBI Databases Follow NCBI NCBI on X (formerly known as Twitter) NCBI on Facebook NCBI on LinkedIn NCBI on GitHub NCBI RSS feed Connect with NLM NLM on X (formerly known as Twitter) NLM on Facebook NLM on YouTube National Library of Medicine 8600 Rockville Pike Bethesda, MD 20894 Web Policies FOIA HHS Vulnerability Disclosure Help Accessibility Careers NLM NIH HHS USA.gov Back to Top

Related documents

• “That level of pressure is... what breaks a pilot”: A qualitative exploration of perceptions around stress and mental health in UK commercial pilots#173883
• Assessing the association between musculoskeletal injury and mental health in rugby players-A cross-sectional study.#145756
• Mental Health in Injured Athletes: A Narrative and Conceptual Review of Risk-Stratified Psychological Monitoring After Musculoskeletal Injury and Concussion.#358165
• Neurobehavioral Changes Following Spaceflight Stressors#1093222
• Integrating health and clinical psychology into sport injury research: A necessary theoretical and applied expansion. A comment on “Psychology of sport injury: Selected debates and contemporary issues” (Podlog and Ivarsson, this issue)#144968