We aimed to use deep learning to detect tuberculosis in chest radiographs in annual workers’ health examination data and compare the performances of convolutional neural networks (CNNs) based on images only (I-CNN) and CNNs including demographic variables (D-CNN). The I-CNN and D-CNN models were trained on 1000 chest X-ray images, both positive and negative, for tuberculosis. Feature extraction was conducted using VGG19, InceptionV3, ResNet50, DenseNet121, and InceptionResNetV2. Age, weight, height, and gender were recorded as demographic variables. The area under the receiver operating characteristic (ROC) curve (AUC) was calculated for model comparison. The AUC values of the D-CNN models were greater than that of I-CNN. The AUC values for VGG19 increased by 0.0144 (0.957 to 0.9714) in the training set, and by 0.0138 (0.9075 to 0.9213) in the test set (both p < 0.05). The D-CNN models show greater sensitivity than I-CNN models (0.815 vs. 0.775, respectively) at the same cut-off point for the same specificity of 0.962. The sensitivity of D-CNN does not attenuate as much as that of I-CNN, even when specificity is increased by cut-off points. Conclusion: Our results indicate that machine learning can facilitate the detection of tuberculosis in chest X-rays, and demographic factors can improve this process.

Kim MS, et al. Ann Occup Environ Med. 2018 Dec;30(1):31. https://doi.org/10.1186/s40557-018-0244-x

The burden of rare diseases on society and patients’ families has increased in Korea. However, because of the infrequency of rare diseases, there is a lack of resources and information to address these cases and inadequate funding for the management of these patients. We investigated the average annual cumulative incidence of rare diseases and the trends in annual cumulative incidence from 2011 to 2015 in Korea by using nationwide administrative data from the Korean National Health Insurance Service (NHIS) database for patients registered with the co-payment assistance policy for rare and incurable diseases. Annual cumulative incidence per 10,000,000 was calculated as the total number of newly enrolled patients with the Korean Standard Classification of Diseases (KCD)-7 code in the register, divided by the number of residents with health insurance coverage during each year. We employed simple linear regression analysis to evaluate the trends in annual cumulative incidence/10,000,000 population per year for each rare disease. Overall, national support was provided for patients with 415 KCD codes listed among the targeted rare diseases. The total number of newly enrolled patients with rare diseases was 53,831 in 2011, 52,658 in 2012, 52,955 in 2013, 71,530 in 2014, and 70,559 in 2015. The number of rare diseases with an average annual cumulative incidence of 100/10,000,000 and above was 22 (5.30%), while there were 227 (54.70%) and 148 (35.66%) with an average cumulative incidence between 1/10,000,000 and 100/10,000,000 and less than 1/10,000,000, respectively. The trends in the annual cumulative incidence for 43 rare diseases were statistically significant (p-value < 0.05). The rare diseases for which the incremental trend per year was statistically significant were sarcoidosis (D86, D86.0, D86.1, D86.2, D86.3, D86.8, D86.9), Parkinson’s disease (G20), Guillain-Barré syndrome (G61.0), primary biliary cirrhosis (K74.3) and Sjogren’s syndrome (M35.0). The number of rare diseases showing an increasing trend in annual cumulative incidence was higher than the number of diseases showing a decreasing trend in annual cumulative incidence. Given that the definition and diagnosis vary based on country and that there is difficulty in identifying valid cases, further detection strategies are needed to establish the incidence of each rare disease considering the importance of establishing a health policy based on the actual incidence of the targeted diseases.

Noise, or undesirable sound, is one of the most common environmental stressors, and it can cause various health effects. Beyond the auditory consequences of occupational noise exposure, extra-auditory effects such as psychological problems have also been found. The aim of the current study is to elucidate the association between occupational noise annoyance and psychological symptoms, including symptoms of depression and suicidal ideation. A total of 10,020 participants (5,410 men and 4,610 women) were included in the current analysis, using data from the fourth Korean National Health and Nutrition Examination Survey (KNHANES). Self-report questionnaires were used to assess noise annoyance levels, depressive symptoms, and suicidal ideation. Odds ratios (ORs) and 95% confidence intervals (95% CIs) for psychosocial symptoms were calculated using multiple logistic regression models. Compared to the no noise annoyance group, ORs (95% CI) of the severe annoyance groups were 1.58 (1.12–2.23) and 1.76 (1.29–2.40) in men and 1.49 (1.05–2.11) and 1.41 (1.01–1.97) in women for depressive symptoms and suicidal ideation, respectively. The ORs (95% CI) for severe noise annoyance in those with less than five hours of sleep were 2.95 (1.46–5.96) and 2.05 (1.01–4.16) in men and women, respectively, compared with those with no noise annoyance and a sleep time of more than five hours. Our study shows that occupational noise annoyance is significantly related to mental health, including depressive symptoms and suicidal ideation after controlling for individual and socio-demographic characteristics even with gender stratification. However, prospective studies with quantified noise exposure assessment were needed to elucidate the causality on the association between noise annoyance and psychological symptoms.