“Nursing necessity” was developed by Tsutsui et al. as an evaluation criterion aimed at reflecting the volume and quality of nursing services in patient-to-nurse ratios. In the 2008 revision of the medical payment system, the evaluation of nursing necessity became obligatory owing to the seven to one patient-to-nurse ratio requirement. However, challenges include the high burden of education and input duties in the evaluation of nursing necessity; moreover, although it is possible to calculate past nursing necessity, future nursing necessity cannot be calculated and used as data that justify the deployment of nurses. By the way, the volume of nursing may be regulated with reference to the order information in electronic medical records by each patient’s condition. Therefore, if nursing necessity could be automatically predicted based on this order, it would be possible to reduce the burden of education and input duties in the evaluation of nursing necessity, and the appropriate deployment of nurses would be possible. Thus, in this study, we attempted to verify the prediction of nursing necessity items based on order data using supervised machine learning. Prediction results indicated the accuracy rate of the nursing necessity items was 0.90 or more for almost all items, and the f value varied from 0-0.85. Furthermore, when predicting the deployment of nurses, the accuracy rate was 0.90 or more for all and number of nurses required is almost same between by nursing necessity and prediction of this method. Therefore, a highly accurate prediction was possible.

Background: The development of speech-recognition systems has progressed considerably, and introducing the system is expected to improve the efficiency of nursing documentation. The clinical research focus has shifted to the usability of documentation systems and knowledge-based discrimination of spoken words. However, differences between spoken and written words could create a bottleneck effect when using speech recognition. To consider the impact of the bottleneck, focusing on circulating nurses, we investigated the extent to which their spoken words corresponded to their nursing records.

Methods: For our investigation we opted for a gynecological laparoscopic surgery which is usually an elective operation. We obtained spoken and nursing record data (hereafter referred to as voice data and report data, respectively) from three surgeries. We first conducted a morphological analysis, and then analyzed the noun morphemes. The noun morphemes from the report data were classified into categories and subcategories, and those from the voice data were compared with the corresponding categories.

Results & Discussion: We obtained 9,220 and 552 morphemes from the voice and report data, of which 2,370 and 450 were noun morphemes, respectively. Within the noun morphemes from the report data, 26.2% were exactly the same as those in the voice data. 63.2% of the noun words in the report data were rephrased in different terms in the voice data. 10.5% of the noun words appeared only in the report data and not in the voice data. We found a high correspondence between voice data and report data. This study demonstrates the possibility of extracting keywords to create a surgical nursing record from the spoken words of circulating nurses. However, many instances of rephrasing were observed. These data suggest the necessity of linking terms used in utterances and documentation.

It is reported that 58% of workers feel severe stress while working. Mechanisms of stress are related to various factors and it was difficult to study them with conventional statistical methods because of multicollinearity. We calculated importance of 31 factors to perceived stress multilaterally with data of a big company’s data-health project. 31 factors included job stressors (branch, commute hours, job description, working hours, midnight work, work life balance, psychological safety, job demand, job control), individual factors (age, occupation, seniority, living with someone, sense of coherence (SOC; to measure stress coping ability)), non-work factors (BMI, smoking status, drinking status, exercise status, housework hours, nursing status, Pittsburgh Sleep Quality Index (PSQI; quality of sleep)), and buffer factors (supervisor support, co-worker support). A self-administered questionnaire was conducted on 7,255 male full-time employees who worked for a certain logistics company. In total, 6,166 (85.0%) of the employees agreed to participate. After we confirmed multivariate configuration factor by principle component analysis, we analyzed the data use a machine learning method, ensemble learning, which is based on a decision tree algorithm. We compared the performance of different algorithms, such as RandomForest (RF), eXtreme Gradient Boosting (XGBoost), Light Gradient Boosting Machine (LightGBM), and Category Boosting (CatBoost). After adjusting the learning model to calculate the importance of the explanatory variable to the objective variable, we adopted the SHapley Additive explanation (SHAP) values, which has consistently been used to interpret model findings. The LightGBM method had the highest performance, so we adopted LightGBM. Regarding importance, SOC was the most important and PSQI was second compared with the remaining variables. These findings indicate the importance of measures for improving stress coping ability and improving quality of sleep compared to job stressors like job demand. This study can be used to help prioritize steps for stress reduction in the workplace.