RESUMO
OBJECTIVE: This study proposed a moving average (MA) approach to dynamically process heart rate variability (HRV) and developed aberrant driving behavior (ADB) prediction models by using long short-term memory (LSTM) networks. BACKGROUND: Fatigue-associated ADBs have traffic safety implications. Numerous models to predict such acts based on physiological responses have been developed but are still in embryonic stages. METHOD: This study recorded the data of 20 commercial bus drivers during their routine tasks on four consecutive days and subsequently asked them to complete questionnaires, including subjective sleep quality, driver behavior questionnaire and the Karolinska Sleepiness Scale. Driving behaviors and corresponding HRV were determined using a navigational mobile application and a wristwatch. The dynamic-weighted MA (DWMA) and exponential-weighted MA were used to process HRV in 5-min intervals. The data were independently separated for training and testing. Models were trained with 10-fold cross-validation strategy, their accuracies were evaluated, and Shapley additive explanation (SHAP) values were used to determine feature importance. RESULTS: Significant increases in the standard deviation of NN intervals (SDNN), root mean square of successive heartbeat interval differences (RMSSD), and normalized spectrum of high frequency (nHF) were observed in the pre-event stage. The DWMA-based model exhibited the highest accuracy for both driver types (urban: 84.41%; highway: 80.56%). The SDNN, RMSSD, and nHF demonstrated relatively high SHAP values. CONCLUSION: HRV metrics can serve as indicators of mental fatigue. DWMA-based LSTM could predict the occurrence of the level of fatigue associated with ADBs. APPLICATION: The established models can be used in realistic driving scenarios.
RESUMO
Objectives. Current approaches via physiological features detecting aberrant driving behaviour (ADB), including speeding, abrupt steering, hard braking and aggressive acceleration, are developing. This study proposes using machine learning approaches incorporating heart rate variability (HRV) parameters to predict ADB occurrence. Methods. Naturalistic driving data of 10 highway bus drivers in Taiwan from their daily routes were collected for 4 consecutive days. Their driving behaviours and physiological data during a driving task were determined using a navigation mobile application and heart rate watch. Participants' self-reported data on sleep, driving-related experience, open-source data on weather and the traffic congestion level were obtained. Five machine learning models - logistic regression, random forest, naive Bayes, support vector machine and gated recurrent unit (GRU) - were employed to predict ADBs. Results. Most drivers with ADB had low sleep efficiency (≤80%), with significantly higher scores in driver behaviour questionnaire subcategories of lapses and errors and in the Karolinska sleepiness scale than those without ADBs. Moreover, HRV parameters were significantly different between baseline and pre-ADB event measurements. GRU had the highest accuracy (81.16-84.22%). Conclusions. Sleep deficit may be related to the increased fatigue level and ADB occurrence predicted from HRV-based models among bus drivers.
Assuntos
Condução de Veículo , Humanos , Acidentes de Trânsito , Frequência Cardíaca/fisiologia , Projetos Piloto , Teorema de Bayes , Aprendizado de MáquinaRESUMO
OBJECTIVES: Chronic otitis media is a long-term infection of the middle ear. It is characterized by persistent discharge from the middle ear through a perforated tympanic membrane. It is one of the most common causes of preventable hearing loss, especially in developing countries. Precise estimation of the size of tympanic membrane perforation is essential for successful clinical management. In this study, we developed a smartphone-based application to calculate the ratio of the area of tympanic membrane perforation to the area of the tympanic membrane. Twelve standardized patients and 60 medical students were involved to assess the area of tympanic membrane perforation, in particular, the percentage of perforation size. METHODS: In total, 60 student doctors (including year 5 and year 6 medical students, intern and post-graduate year training of doctors) were recruited during their rotation at the Otolaryngology department of Taipei Medical University Shuang-Ho Hospital. Twelve standardized patients with chronic otitis media were recruited through a single otology practice. Oto-endoscopic examination was performed for all patients by using a commercially-available digital oto-endoscope, and clinical images of the tympanic membrane perforation were obtained. To demonstrate the variability of perforation size estimation by different student doctors, we calculated the percentage of perforation using the smartphone-based application for 12 tympanic membranes objectively and compared the results with those visually estimated by the 60 student doctors subjectively. RESULTS: The variance in the visual estimation by the 60 student doctors was large. By contrast, variances in smartphone-based application calculations were smaller, indicating consistency in the results obtained from different users. The smartphone-based application accurately estimated the presence of perforation for tympanic membranes with high consistency. The differences in visual estimations can be considerably great and the variances can be large among different individuals. CONCLUSIONS: The smartphone-based application is a dependable tool for precisely evaluating the size of tympanic membrane perforation.
RESUMO
(a) Objective: Obstructive sleep apnea syndrome (OSAS) is typically diagnosed through polysomnography (PSG). However, PSG incurs high medical costs. This study developed new models for screening the risk of moderate-to-severe OSAS (apnea-hypopnea index, AHI ≥15) and severe OSAS (AHI ≥30) in various age groups and sexes by using anthropometric features in the Taiwan population.(b) Participants: Data were derived from 10,391 northern Taiwan patients who underwent PSG.(c) Methods: Patients' characteristics - namely age, sex, body mass index (BMI), neck circumference, and waist circumference - was obtained. To develop an age- and sex-independent model, various approaches - namely logistic regression, k-nearest neighbor, naive Bayes, random forest (RF), and support vector machine - were trained for four groups based on sex and age (men or women; aged <50 or ≥50 years). Dataset was separated independently (training:70%; validation: 10%; testing: 20%) and Cross-validated grid search was applied for model optimization. Models demonstrating the highest overall accuracy in validation outcomes for the four groups were used to predict the testing dataset.(d) Results: The RF models showed the highest overall accuracy. BMI was the most influential parameter in both types of OSAS severity screening models.(e) Conclusion: The established models can be applied to screen OSAS risk in the Taiwan population and those with similar craniofacial features.
Assuntos
Apneia Obstrutiva do Sono , Masculino , Humanos , Feminino , Taiwan/epidemiologia , Teorema de Bayes , Polissonografia , Apneia Obstrutiva do Sono/diagnóstico , Apneia Obstrutiva do Sono/epidemiologia , Aprendizado de MáquinaRESUMO
As part of the new measures to prevent the spread of the 2019 coronavirus disease (COVID-19), medical students were advised to wear a mask in class and avoid touching their faces. Few studies have analyzed the influence of health education on the frequency of face- and smartphone-touching behaviors during the COVID-19 pandemic. This research compared the frequency of in-class face- and smartphone-touching behaviors of medical students before and after the delivery of personal hygiene education during the COVID-19 pandemic. A behavioral observational study was conducted involving medical students at Taipei Medical University. Eighty medical students were recruited during a lecture on otorhinolaryngology. All medical students were required to wear a mask. Their face- and smartphone-touching behavior was observed by viewing the 4 k resolution video tape recorded in class. The recording lasted for 2 h, comprising 1 h prior to the health educational reminder and 1 h afterwards. The frequencies of hand-to-face contact and hand-to-smartphone contact were analyzed before and after the delivery of health education emphasizing personal hygiene. Comprehensive health education and reminders effectively reduce the rate of face- and smartphone-touching behaviors.
Assuntos
COVID-19 , Pandemias , Humanos , Higiene , Pandemias/prevenção & controle , SARS-CoV-2 , SmartphoneRESUMO
Hearing impairment is a frequent human sensory impairment. It was estimated that over 50% of those aged >75 years experience hearing impairment in the United States. Several hearing impairment-related factors are detectable through screening; thus, further deterioration can be avoided. Early identification of hearing impairment is the key to effective management. However, hearing screening resources are scarce or inaccessible, underlining the importance of developing user-friendly mobile health care systems for universal hearing screening. Mobile health (mHealth) applications (apps) act as platforms for personalized hearing screening to evaluate an individual's risk of developing hearing impairment. We aimed to evaluate and compare the accuracy of smartphone-based air conduction and bone conduction audiometry self-tests with that of standard air conduction and bone conduction pure-tone audiometry tests. Moreover, we evaluated the use of smartphone-based air conduction and bone conduction audiometry self-tests in conductive hearing loss diagnosis. We recruited 103 patients (206 ears) from an otology clinic. All patients were aged ≥20 years. Patients who were diagnosed with active otorrhea was excluded. Moderate hearing impairment was defined as hearing loss with mean hearing thresholds >40 dB. All patients underwent four hearing tests performed by a board-certified audiologist: a smartphone-based air conduction audiometry self-test, smartphone-based bone conduction audiometry self-test, standard air-conduction pure-tone audiometry, and standard bone conduction pure-tone audiometry. We compared and analyzed the results of the smartphone-based air conduction and bone conduction audiometry self-tests with those of the standard air conduction and bone conduction pure-tone audiometry tests. The sensitivity of the smartphone-based air conduction audiometry self-test was 0.80 (95% confidence interval CI = 0.71-0.88) and its specificity was 0.84 (95% CI = 0.76-0.90), respectively. The sensitivity of the smartphone-based bone conduction audiometry self-test was 0.64 (95% CI = 0.53-0.75) and its specificity was 0.71 (95% CI = 0.62-0.78). Among all the ears, 24 were diagnosed with conductive hearing loss. The smartphone-based audiometry self-tests correctly diagnosed conductive hearing loss in 17 of those ears. The personalized smartphone-based audiometry self-tests correctly diagnosed hearing loss with high sensitivity and high specificity, and they can be a reliable screening test to rule out moderate hearing impairment among the population. It provided patients with moderate hearing impairment with personalized strategies for symptomatic control and facilitated individual case management for medical practitioners.
RESUMO
BACKGROUND: Hearing impairment is the most frequent sensory deficit in humans, affecting more than 360 million people worldwide. In fact, hearing impairment is not merely a health problem, but it also has a great impact on the educational performance, economic income, and quality of life. Hearing impairment is therefore an important social concern. OBJECTIVE: We aimed to evaluate and compare the accuracy of self-perception, Hearing Handicap Inventory for the Elderly-Screening (HHIE-S) questionnaire, free-field voice test, and smartphone-based audiometry as tests for screening moderate hearing impairment in older adults in China. METHODS: In this study, 41 patients were recruited through a single otology practice. All patients were older than 65 years. Patients with otorrhea and cognitive impairment were excluded. Moderate hearing impairment was defined as mean hearing thresholds at 500, 1000, 2000, and 4000 Hz >40 dB hearing loss (pure-tone average > 40 dB hearing loss). All patients completed 5 hearing tests, namely, the self-perception test, HHIE-S questionnaire test, free-field voice test, smartphone-based audiometry test, and standard pure-tone audiometry by the same audiologist. We compared the results of these tests to the standard audiogram in the better-hearing ear. RESULTS: The sensitivity and the specificity of the self-perception test were 0.58 (95% CI 0.29-0.84) and 0.34 (95% CI 0.19-0.54), respectively. The sensitivity and the specificity of the HHIE-S questionnaire test were 0.67 (95% CI 0.35-0.89) and 0.31 (95% CI 0.316-0.51), respectively. The sensitivity and the specificity of the free-field voice test were 0.83 (95% CI 0.51-0.97) and 0.41 (95% CI 0.24-0.61), respectively. The sensitivity and the specificity of the smartphone-based audiometry test were 0.92 (95% CI 0.60-0.99) and 0.76 (95% CI 0.56-0.89), respectively. Smartphone-based audiometry correctly diagnosed the presence of hearing loss with high sensitivity and high specificity. CONCLUSIONS: Smartphone-based audiometry may be a dependable screening test to rule out moderate hearing impairment in the older population.