Diabetic complication prediction using a similarity-enhanced latent Dirichlet allocation model

Identifying Diabetes Related-Complications in a Real-World Free-Text Electronic Medical Records in Hebrew Using Natural Language Processing Techniques

BACKGROUND

Studies have demonstrated that 50% to 80% of patients do not receive an International Classification of Diseases (ICD) code assigned to their medical encounter or condition. For these patients, their clinical information is mostly recorded as unstructured free-text narrative data in the medical record without standardized coding or extraction of structured data elements. Leumit Health Services (LHS) in collaboration with the Israeli Ministry of Health (MoH) conducted this study using electronic medical records (EMRs) to systematically extract meaningful clinical information about people with diabetes from the unstructured free-text notes.

OBJECTIVES

To develop and validate natural language processing (NLP) algorithms to identify diabetes-related complications in the free-text medical records of patients who have LHS membership.

METHODS

The study data included 2.3 million records of 41 469 patients with diabetes aged 35 or older between the years 2012 and 2017. The diabetes related complications included cardiovascular disease, diabetic neuropathy, nephropathy, retinopathy, diabetic foot, cognitive impairments, mood disorders and hypoglycemia. A vocabulary list of terms was determined and adjudicated by two physicians who are experienced in diabetes care board certified diabetes specialist in endocrinology or family medicine. Two independent registered nurses with PhDs reviewed the free-text medical records. Both rule-based and machine learning techniques were used for the NLP algorithm development. Precision, recall, and F-score were calculated to compare the performance of (1) the NLP algorithm with the reviewers' comments and (2) the ICD codes with the reviewers' comments for each complication.

RESULTS

The NLP algorithm versus the reviewers (gold standard) achieved an overall good performance with a mean F-score of 86%. This was better than the ICD codes which achieved a mean F-score of only 51%.

CONCLUSION

NLP algorithms and machine learning processes may enable more accurate identification of diabetes complications in EMR data.

Tracing the evolution of green logistics: A latent dirichlet allocation based topic modeling technology and roadmapping

Green logistics (GL) is the main development trend of modern logistics. The analysis of green logistics topics and their evolution is helpful in grasping its development trend and doing research facing the international frontier. Focusing on the hot topics and evolution process of green logistics, this paper analyzes from four aspects: firstly, this study divides the green logistics development progress into three stages based on life cycle theory, which are the emerging stage (1993-2003), slow growth stage (2004-2014) and rapid growth stage (2015-2021). Then, based on latent dirichlet allocation (LDA) topic model, this study summarizes and confirms related words and meaning of each topic in different stages. Furthermore, this study calculates the text similarity in each development stage of green logistics and analyzes the trend of hot topics in green logistics. Finally, this paper visualizes the development roadmap of green logistics and explores the progression among three stages. There are 4, 5, and 7 topics defined respectively in three development stages. The revolution of green logistics is analyzed, and the results show that "model and management on sustainable development of GL", "related issues and potential threats of GL", and "optimization analysis of low-carbon vehicle routing and time" are the primary development trends of green logistics. This study fills the gap in considering the evolution process of green logistics through topic modeling and roadmapping method. It provides a particular theoretical significance for the green and sustainable development of logistics.

Role of vaccine efficacy in the vaccination behavior under myopic update rule on complex networks

How to effectively prevent the diffusion of infectious disease has become an intriguing topic in the field of public hygienics. To be noted that, for the non-periodic infectious diseases, many people hope to obtain the vaccine of epidemics in time to be inoculated, rather than at the end of the epidemic. However, the vaccine may fail as a result of invalid storage, transportation and usage, and then vaccinated individuals may become re-susceptible and be infected again during the outbreak. To this end, we build a new framework that considers the imperfect vaccination during the one cycle of infectious disease within the spatially structured and heterogeneous population. Meanwhile, we propose a new vaccination update rule: myopic update rule, which is only based on one focal player's own perception regarding the disease outbreak, and one susceptible individual makes a decision to adopt the vaccine just by comparing the perceived payoffs vaccination with the perceived ones of being infected. Extensive Monte-Carlo simulations are performed to demonstrate the imperfect vaccination behavior under the myopic update rule in the spatially structured and heterogeneous population. The results indicate that healthy individuals are often willing to inoculate the vaccine under the myopic update rule, which can stop the infectious disease from being spread, in particular, it is found that the vaccine efficacy influences the fraction of vaccinated individuals much more than the relative cost of vaccination on the regular lattice, Meanwhile, vaccine efficacy is more sensitive on the heterogeneous scale-free network. Current results are helpful to further analyze and model the choice of vaccination strategy during the disease outbreaks.