Search

Background: Burnout is a significant public health concern affecting more than half of the healthcare workforce; however, passive screening tools to detect burnout are lacking. We investigated the ability of machine learning (ML) techniques to identify burnout using passively collected electronic health record (EHR)-based audit log data.Method: Physician trainees participated in a longitudinal study where they completed monthly burnout surveys and provided access to their EHR-based audit logs. Using the monthly burnout scores as the target outcome, we trained ML models using combinations of features derived from audit log data-aggregate measures of clinical workload, time series-based temporal measures of EHR use, and the baseline burnout score. Five ML models were constructed to predict burnout as a continuous score: penalized linear regression, support vector machine, neural network, random forest, and gradient boosting machine.Results: 88 trainee physicians participated and completed 416 surveys; greater than10 million audit log actions were collected (Mean [Standard Deviation] = 25,691 [14,331] actions per month, per physician). The workload feature set predicted burnout score with a mean absolute error (MAE) of 0.602 (95% Confidence Interval (CI), 0.412-0.826), and was able to predict burnout status with an average AUROC of 0.595 (95% CI 0.355-0.808) and average accuracy 0.567 (95% CI 0.393-0.742). The temporal feature set had a similar performance, with MAE 0.596 (95% CI 0.391-0.826), and AUROC 0.581 (95% CI 0.343-0.790). The addition of the baseline burnout score to the workload features improved the model performance to a mean AUROC of 0.829 (95% CI 0.607-0.996) and mean accuracy of 0.781 (95% CI 0.587-0.936); however, this performance was not meaningfully different than using the baseline burnout score alone.Conclusions: Current findings illustrate the complexities of predicting burnout exclusively based on clinical work activities as captured in the EHR, highlighting its multi-factorial and individualized nature. Future prediction studies of burnout should account for individual factors (e.g., resilience, physiological measurements such as sleep) and associated system-level factors (e.g., leadership). [ABSTRACT FROM AUTHOR]

This study investigated the wintertime vertical distributions and source areas of aerosols, NO 2 , and HCHO in a coastal city of Dongying from December 2020 to March 2021, using ground-based multi-axis differential optical absorption spectroscopy (MAX-DOAS) and a potential source contribution function (PSCF) model, respectively. Moreover, the chemical production sensitivity of O 3 at different height layers was analyzed using HCHO/NO 2 ratios. The results revealed that the wintertime averaged highest concentrations of aerosol (1.25 km−1), NO 2 (14.81 ppb), and HCHO (2.32 ppb) were mainly distributed at the surface layer, 100–200 m layer, and 200–300 m layer, respectively. Regarding the diurnal cycles, high concentrations of aerosol (>1.4 km−1) and NO 2 (>16.0 ppb) usually appeared in the early morning and late afternoon, while high concentrations of HCHO (>2.5 ppb) usually occurred during 12:00–15:00. The PSCF model revealed that the wintertime aerosol mainly originated from Shandong, northern Jiangsu, Korea, and the northwestern Mongolian Plateau. Below 200 m, NO 2 was mainly from western Shandong, whereas above 600 m, it was mainly from northern Shandong and the Beijing-Tianjin-Hebei (BTH) region. The corresponding sources for HCHO were central and southern Shandong (below 200 m) and northern Shandong, northern Jiangsu, and southeastern BTH (above 600 m). In addition, the chemical production sensitivity of O 3 below 100 m was observed only in the VOC-limited regime. The percentages of O 3 production under the NO x -limited, NO x -VOC-limited, and VOC-limited regimes were 10.75% (31.18%), 4.30% (19.35%), and 84.95% (49.47%) at the 500–600 m (900–1000 m) layer. This study has guiding significance for the coordinated control of PM 2.5 and O 3 , and can assist in the implementation of regional joint prevention and control strategies for air pollutants. • The averaged wintertime NO 2 and HCHO profiles all exhibited a Gaussian shape. • The vertical potential sources of aerosol, NO 2 and HCHO were analyzed using PSCF. • O 3 chemical production sensitivity under 100 m was only VOC-limited. • The percentages of VOC-limited O 3 production at 500–600 were 84.95%. [ABSTRACT FROM AUTHOR]

• A review of blind deconvolution methods (BDMs) in machinery fault diagnosis. • History, principle, classic method and applications of BDMs are reviewed. • Modified BDMs are classified to review according to their basic principle and merits and limitations are verified. • Future prospects of BDMs in machinery fault detection is discussed. Fault diagnosis is of significance for ensuring the safe and reliable operation of machinery equipment. Due to the heavy noise and interference, it is difficult to detect the fault directly from the measured signal. Hence, signal processing techniques that can achieve feature extraction, signal denoising, and fault identification are the most common tools in the field. Blind deconvolution methods (BDMs), as one of the most classic methods, have been studied extensively and applied fully for machinery fault diagnosis. Up to now, plenty of publications about the studies and applications of BDMs for machinery fault diagnosis have been presented to academic journals, technical reports, and conference proceedings. This paper intends to survey and summarize the current progress of BDMs applied in machinery fault diagnosis, as well as provides a comprehensive review of BDMs from history to state-of-the-art methods and finally to research prospects. Firstly, the theoretical background and brief history of BDMs are introduced. Secondly, the modified BDMs are classified to review their basic principles. After that their merits and limitations as well as the performance analysis are summarized. Thirdly, the research and application on machinery fault detection using BDMs are overviewed. Finally, the prospects of BDMs in machinery fault diagnosis are discussed. [ABSTRACT FROM AUTHOR]